Hepatic arterial infusion PUMP chemotherapy combined with systemic therapy versus systemic therapy alone as Induction Therapy for initially unresectable colorectal liver metastases: a randomised controlled trial

In patients with colorectal cancer metastases confined to the liver (CRLM),
resection of the primary tumour and liver metastases may offer a chance of cure
or long-term survival. The vast majority of these patients however presents
with unresectable CRLM. Resection is a dominant contributor to long-term
survival. Therefore, major efforts have focused on optimising induction
treatment to convert unresectable CRLM to resectable CRLM.
Induction treatment with systemic therapy is the current treatment regime with
reported conversion rates ranging from 22 to 57% due to heterogeneity of the
patient population and variation in resectability criteria. The recently
published Dutch CAIRO5 study was designed to find the optimal induction regimen
in chemotherapy-naive patients with initially unresectable CRLM based on
predefined resectability criteria. Patients with right-sided and/or
RAS/BRAFV600E mutated primary tumours were randomised to FOLFOX (5FU,
leucovorin and oxaliplatin) or FOLFIRI (5FU, leucovorin and irinotecan) +
bevacizumab (Bev) (arm A) and FOLFOXIRI (5FU, leucovorin, oxaliplatin and
irinotecan) + Bev (arm B). Patients with left-sided and RAS/BRAFV600E wild-type
tumours were randomised to FOLFOX/FOLFIRI (patient preference) plus either
bevacizumab(arm C) or panitumumab (arm D). Besides a higher conversion rate
after triplet therapy (arm B), no difference in progression free survival and
overall survival was found. This could partially be explained by the high
recurrence rates after conversion surgery. Early recurrence, defined as disease
progression or death within 6 months after conversion, occurred in 43% of 240
patients who underwent conversion surgery. The majority (62%) had liver-only
recurrence. Considering that the liver is the predominant site of recurrences,
intensifying induction treatment might improve survival outcomes.
Hepatic arterial infusion pump (HAIP) chemotherapy is a promising method for
optimising induction treatment. The HAIP is surgically placed and delivers
chemotherapy continuously and directly to the liver via the hepatic artery.
Floxuridine (FUDR) is the chemotherapeutic agent used in HAIP therapy. FUDR has
a high first-pass liver extraction rate of 95% which results in limited
systemic side effects and enables a 400-fold higher dose compared to systemic
administration. In patients with unresectable CRLM, HAIP therapy is combined
with systemic therapy (HAIP-SYST) to optimise conversion rates and control
occult extrahepatic disease. The reported conversion rates after HAIP-SYST vary
from 18-52%. In line with other studies describing induction therapy for
unresectable CRLM, the wide range in conversion rates after HAIP-SYST could be
explained by the heterogenity of both resectability criteria and the patient
population.

In this diverse patient population, chemotherapy-naive patients appear to
benefit most from HAIP-SYST as induction treatment. They show improved survival
outcomes, in particular patients who convert to local treatment of the liver
metastases. We analysed 58 consecutive chemo-naive patients with unresectable
CRLM who were treated with HAIP-SYST at Memorial Sloan Kettering Cancer Center
(MSKCC). Thirty-two patients (55%) converted to resectable CRLM and another two
patients had a radiological complete response. After median follow-up of 116
months, median PFS was 16.7 months and median OS was 53.0 months for all
patients. In converted patients or those with a radiological complete response
versus patients who remained unresectable, the 3 year OS rate was 88% vs. 27%,
and the 5 year OS rate was 72% vs. 0% (P <0.001). Moreover, 19 converted
patients (59%) had a complete or major pathological response.

These results of HAIP-SYST in chemo-naive patients are impressive, but until
now just reported from a single centre. Most important barriers for
implementing HAIP-SYST treatment worldwide include absence of marketing
authorization for FUDR in Europe, the technically challenging surgical
procedure of pump placement and the need for stringent monitoring and specific
management of patients undergoing HAIP-SYST which requires a highly skilled
multidisciplinary treatment team.
To overcome these barriers we first conducted a phase II study in the
Netherlands Cancer Institute and Erasmus MC to assess safety and feasibility of
HAIP-SYST as induction treatment. The endpoint was feasibility, described as
the percentage of patients completing 2 cycles of combined HAIP chemotherapy
and systemic therapy. This endpoint was reached in 28 of 31 (90%) patients.
Pump placement was combined with resection of the primary tumour in 22 of 30
patients. Anastomotic leakage and mortality were not observed. Clavien-Dindo
grade >= 3 complications of the pump placement procedure (± resection of the
primary tumour) were documented in 5 of 31 (16%) patients. Median time from
surgery to first HAIP was 21 days (IQR 15-28). CTCAE toxicity grade 3 of more
was seen in 10 of 29 (23%) patients during the first two courses of HAIP-SYST.
After proven save and feasible, this multicentre randomised controlled trial is
the subsequent step to provide insights on long-term outcomes of combined
HAIP-SYST as induction treatment over systemic therapy alone for
chemotherapy-naive patients with unresectable CRLM.

The main objective of this trial is to investigate whether HAIP with
concomitant systemic therapy prolongs overall survival in chemo-naive patients
with initially unresectable synchronous colorectal liver metastases as compared
with systemic therapy alone.

Multicentre, phase 3 randomised controlled trial

Patients in the intervention arm will first undergo surgery for pump placement
and resection of the primary tumour. Subsequently, these patients receive
induction therapy with FUDR-HAIP combined with systemic therapy (FOLFOX or
FOLFIRI). The control arm receives systemic therapy alone according to standard
clinical practice (FOLFOX/FOLFIRI/CAPOX/FOLFOXIRI (+bevacizumab)).
A National Liver Panel consisting of radiologists and liver surgeons assess the
CT-abdomen of all patients at baseline and during induction treatment. The
panel evaluates resectability and treatment response. The duration of protocol
treatment depends on toxicity, disease progression and treatment response. In
case of conversion to resectable CRLM, the total duration of treatment is 6
months (i.e. 6 cycles) of active treatment: HAIP-SYST or SYST.
Patients in the intervention arm need an additional preoperative CT angiography
to assess suitability for pump catheter placement and postoperative 99mTc-MAA
scintigraphy to exclude extrahepatic perfusion.
All patients will be asked to complete quality of life questionnaires at
baseline and 3, 6, 9, 12 and 24 months after starting systemic therapy or after
pump placement.

The anticipated benefit of HAIP-SYST induction therapy is longer overall
survival, higher cure rate and better quality of life. Conversion to local
treatment of CRLM with curative intent or a durable complete radiological
response has been observed in 59% of chemo-naive patients who were treated with
a combination of HAIP chemotherapy and first line systemic therapy.
Irrespective of conversion, the combined therapy results in a median overall
survival (OS) of 53 months and 5-year OS in 4% of patients. In case of
conversion-to-surgery or complete radiological response, median OS was 141.7
months and 5-year OS was 72%. These data are impressive but retrospectively
derived from single centre experience. Conversion rates and survival benefit
have not yet been established elsewhere. We anticipate that a higher conversion
rate will result in time without chemotherapy and therewith better quality of
life. We anticipate that the cure rate, which is generally very low in this
patient population, will be higher in the intervention arm because
micrometastatic hepatic disease is eradicated in more patients.
In patients randomised to the intervention arm, HAIP will be combined with
systemic therapy. To assess arterial anatomy, a CT scan with arterial phase has
to be performed if not already available. The chemo pump will be implanted
surgically and resection of the primary tumour will be performed during the
same procedure. Surgical complications related to chemo pump implantation are
uncommon (<10%) and include hepatic artery bleeding, dissection or thrombosis,
HAIP pocket infection, and extrahepatic perfusion. Complications associated
with surgery in general and colorectal surgery specifically may occur. These
complications include bleeding, infection, or anastomotic leak. Due to the
mandatory surgical procedure and recovery period, there is usually a short
delay in start of chemotherapy. Surgical complications can lead to
re-interventions and mortality. Severe complications may significantly postpone
systemic treatment, or even keep the patient from treatment.
Prior to the first administration of HAIP chemotherapy, a
technetium-99m-labeled macro-aggregated albumin nuclear medicine scan
(99mTc-MAA scintigraphy) will be performed to confirm bilobar hepatic perfusion
via the pump and to rule out extrahepatic perfusion. The effective radiation
dose of the 99mTc-MAA scintigraphy is 3-4 mSv.
HAIP chemotherapy toxicities include ulcer disease and biliary sclerosis, which
can both be largely avoided by imaging prior to treatment, monitoring of liver
tests and dose adjustments (or cessation) or concurrent administration of
dexamethasone via pump. Systemic side effects of HAIP chemotherapy are rare
(<1%). Therefore this treatment is suitable for combined induction treatment.
The systemic therapy administered in this trial is well known standard
treatment in metastatic colorectal cancer. Common toxicities include fatigue,
nausea, diarrhoea, neutropenia, palmar-plantar erythrodysesthesia (for 5-FU
only), neuropathy (for oxaliplatin only) and alopecia (for irinotecan only).
When combined with FUDR-HAIP, the systemic chemotherapy dose is slightly
reduced compared to the standard schemes. mFOLFOX6 is modified to a dose of
2000 mg/m2 fluorouracil without bolus of fluorouracil vs. 2400 mg/mg2
fluorouracil, respectively. FOLFIRI is modified by reducing the irinotecan dose
is from 180 mg/m2 to 150 mg/m2 to reduce the risk of biliary toxicity.
Visits to the day care unit have a similar frequency for patients in the
intervention arm and the control arm. Patients in both arms have planned visits
every two weeks for systemic chemotherapy and patients in the intervention arm
will additionally receive filling of the pump either with saline or FUDR.
Participating patients are asked to fill out Quality of Life questionnaires at
baseline after inclusion, and 3, 6, 9, 12 and 24 months after the start of
systemic therapy or after pump placement, in the control arm and intervention
arm respectively.