Background
The epidermal growth factor receptor (EGFR) is a cell surface protein that has become an important therapeutic target in colorectal cancer (CRC) [
1]. Two monoclonal antibodies that target the extracellular domain of the EGFR have been developed: cetuximab (Erbitux), which is a recombinant immunoglobulin G1 mouse–human chimeric anti-EGFR monoclonal antibody (mAb), and panitumumab (Vectibix), a recombinant, fully human immunoglobulin G2 anti-EGFR mAb [
1,
2]. Anti-EGFR therapy (treatment with cetuximab or panitumumab) has been shown to be effective in metastatic CRC (mCRC) [
3‐
6]. Initially, patients with tumours that had mutations of exon 2 of the
KRAS oncogene were found to be resistant to treatment with anti-EGFR mAbs [
3,
7,
8]. Further studies provided evidence that additional mutations beyond
KRAS exon 2 occurring in the wider
RAS family of oncogenes, specifically in exons 3 and 4 of
KRAS and exons 2, 3 and 4 of
NRAS, are also predictive of a lack of response to anti-EGFR therapy [
9‐
13].
Panitumumab was first approved in the European Union (EU) in December 2007 as monotherapy for the treatment of patients with mCRC and confirmed wild-type
KRAS tumour status after failure of fluoropyrimidine-, oxaliplatin- or irinotecan-containing chemotherapy [
3,
14]. However, in November 2011, the approved licence for panitumumab was extended to cover its use as a first-line agent in combination with 5-fluorouracil/folinic acid + oxaliplatin (FOLFOX) chemotherapy and as second line in combination with 5-fluorouracil/folinic acid + irinotecan (FOLFIRI) chemotherapy, again restricted to patients with confirmed wild-type
KRAS tumour status [
7,
15]. In June 2013, EU treatment guidelines changed to recommend that panitumumab should be prescribed to patients with mCRC and wild-type
RAS tumour status (exons 2, 3, 4 of
KRAS and
NRAS), which should be confirmed prior to treatment initiation [
16,
17]. The current label for panitumumab includes a contraindication for its use in combination with oxaliplatin-containing chemotherapy in patients with mutant or unknown
RAS tumour status (or
KRAS status before June 2013) [
16].
Two studies – a physician survey and a medical records review (MRR) – were initiated in Europe in 2012 to evaluate physicians’ awareness of the correct therapeutic indication for panitumumab and to establish if it was being prescribed in accordance with this indication, which is to patients with mCRC and confirmed wild-type
KRAS tumour status prior to treatment with panitumumab. The studies were carried out in three consecutive rounds; the results of the first two rounds of both studies have been published previously [
18]. Overall in Rounds 1 and 2 of the physician survey, 298 (99.0%) of 301 physicians responded correctly that panitumumab should be administered only to patients with confirmed
KRAS wild-type tumours. In Rounds 1 and 2 of the MRR study, 299 (97.7%) of 306 patients reportedly had confirmed wild-type
KRAS status before the initiation of panitumumab treatment. Of 85 patients who were prescribed panitumumab with concurrent oxaliplatin-containing chemotherapy in Rounds 1 and 2 of the MRR, 83 (97.6%) had confirmed wild-type
KRAS status before the initiation of treatment [
18].
The results of the third round of the physician survey and MMR, which focused on full RAS testing, are presented here. The primary objective of Round 3 of the physician survey was to assess physicians’ knowledge of the updated indication for panitumumab treatment, following the changes to the label modifying its use from KRAS wild-type to RAS wild-type mCRC only. Similarly, for Round 3 of the MRR, the main aim was to estimate the prevalence of full RAS testing in the routine clinical management of patients being prescribed panitumumab.
Methods
Physician survey and MRR overview
The detailed methodology of Rounds 1 and 2 (assessing
KRAS only) for both the physician survey and MRR, conducted from 2012 to 2013, has been published in full previously [
18].
Rounds 3 of the physician survey and the MRR were carried out from September 2014 to November 2014 and September 2014 to June 2015, respectively. Physicians from the following nine European countries were invited to participate in the studies: Belgium, Czech Republic, Denmark, France, Germany, Italy, Spain, the Netherlands and Sweden.
For Round 3 of both studies, the following data sources were used to select a random sample of potential participants: (a) a medical marketing database provided by a healthcare industry provider (Cegedim S.A., Boulogne-Billancourt, France), filtered by specialty, and (b) a list of CRC physicians provided by local affiliates of the study sponsor (Amgen Ltd., Uxbridge, UK). Round 3 of the MRR was carried out by Amgen Ltd. and Round 3 of the physician survey was carried out by a separate industry provider (Adelphi Research, Bollington, UK).
Eligibility criteria for the physician survey
Practising oncologists were included in Round 3 of the physician survey if they had treated at least three new or continuing patients with mCRC in the 3 months immediately preceding their invitation to participate in the survey, and only if they had prescribed panitumumab at least once during the previous 6 months. Potential participants were excluded if they had previously taken part in either Round 1 or Round 2 of the survey.
Eligibility criteria for the MRR
Practising oncologists were included in Round 3 of the MRR if they had treated at least three new or continuing patients with mCRC in the 3 months immediately prior to receiving their invitation to participate. In addition, oncologists were only eligible if they had prescribed a first dose of panitumumab to treat a new patient with mCRC in the preceding 6 months. Again they were excluded if they had already taken part in either Round 1 or Round 2 of the MRR. In addition, only one oncologist per participating medical centre was permitted to participate in the MRR, in each round of the study.
Patients were eligible to be included in the MRR by their oncologist if they had received their first dose of panitumumab during the 6-month period before the time at which medical records were accessed for the study. As with participating oncologists, patients were excluded from the MRR if they had taken part in Rounds 1 or 2. Patients were also excluded if they were participating in an experimental clinical trial at the time of receiving panitumumab.
Data collection
In Round 3 of the physician survey, telephone interviews were conducted with eligible oncologists using a standardised questionnaire (see Additional file
1 for interview guide) and following consistent data-collection procedures.
For each oncologist included in the MRR, the relevant anonymised information for eligible patients who had received their first dose of panitumumab was abstracted from their medical records using standardised forms. The oncologists were also asked to identify the pathology centre that performed the RAS (or KRAS) testing. Further information was then collected and reported by the pathologists at these centres, again using a standardised questionnaire.
Statistical analysis
The data analysis was descriptive for both the physician survey and MRR. For the categorical study endpoints, the count and proportion (%) in each category, based on the appropriate denominator, were calculated. The 95% confidence intervals were calculated for the proportions based on a normal approximation to the binomial distribution.
Discussion
Recent changes to the prescribing guidelines for anti-EGFR mAbs (panitumumab and cetuximab) require
RAS tumour genotyping to be carried out for patients with mCRC prior to the initiation of these therapies. These revisions have highlighted the need to gain a better understanding on prescribing oncologists’ awareness of these changes. The physician survey and MRR, the third rounds of which have been described here, were carried out to assess physicians’ knowledge of the updated indication for panitumumab treatment, following the changes to the label from
KRAS to
RAS mutation testing [
18].
In Round 3 of the physician survey, all oncologists who participated were aware that
RAS testing should be performed before their patients’ first dose of panitumumab. Further to this, nearly all of the oncologists (95.4%) also correctly identified that panitumumab is indicated for the treatment of mCRC in patients with confirmed wild-type
RAS tumour status. These findings are consistent with the results of Rounds 1 and 2 of the physician survey, conducted in 2012–2013 before the latest label changes for panitumumab, where the majority (99.0%) of participants correctly identified that
KRAS testing should be performed in patients with mCRC and confirmed wild-type
KRAS tumours, in accordance with the then-correct indication for panitumumab [
18].
In Round 3 of the MRR, all of the patients whose medical records were investigated had a confirmed wild-type tumour status before the initiation of panitumumab treatment, although a minority (16.8%) were only tested for KRAS mutation status. All patients included in the MRR who were treated with panitumumab and concurrent oxaliplatin-containing chemotherapy had confirmed wild-type tumour status, again though a small number (9.9%) were only tested for KRAS mutation status.
Nineteen of the oncologists (12.5%) who participated in Round 3 of the physician survey confirmed that they had administered panitumumab to at least one patient with mCRC and mutant or unknown RAS tumour status within the 6 months prior to completing the survey. The reasons given for these treatment decisions indicate that there may be clinical considerations relating to a patient’s clinical status, the practicalities of RAS testing, or the possibility that in later lines of therapy, patients and physicians may resort to treatments which are either not included in, or deviate from, guidelines. This suggests there is still a need for physician education which would enable each of these obstacles to be easily overcome and lead to improved practice so that all mCRC patients have a confirmed wild-type RAS tumour status before starting treatment with panitumumab.
In contrast to the physician survey, Round 3 of the MRR found that all patients studied had a confirmed wild-type tumour status prior to the initiation of panitumumab treatment; however, this was accounting for both
KRAS and
RAS testing, and the former was not explicitly asked about in the physician survey, which may in part explain the disparity. Furthermore, the physician survey and MRR results are not directly comparable, due to differences in the question regarding off-label prescription of panitumumab (the physician survey assessed the percentage of physicians who prescribed off-label to at least one patient in the last 6 months, and the MMR assessed the percentage of patients who were prescribed off-label panitumumab). These results were, again, broadly consistent with the combined results from Rounds 1 and 2 of both studies, which found that 5.0% of oncologists surveyed had treated a patient with panitumumab when they had either an unknown or mutant
KRAS status while the MRR found that only 2.3% of patients had received panitumumab without having a confirmed wild-type
KRAS status [
18].
As other studies have shown, a minority of laboratories in Europe have continued to use
KRAS testing since June 2013, despite the panitumumab label change [
19]. This is important to note, both because of the update to the indication for anti-EGFR therapies and also because
KRAS mutation testing is less sensitive than full
RAS testing [
20]. However, a further examination of the data from Round 3 of the MRR identified that 18 of the 22 samples tested for
KRAS only had a test report date before the start of 2014, suggesting that they may have been carried out either before or immediately after the change to the prescribing guidelines. Additionally, in Round 3, each tumour sample was classified as having been tested for either
RAS or
KRAS based exclusively on the information recorded in the oncologist notes; the classification was not based on the specific exons and codons tested by the pathologist as this is often not recorded. As this is information which could not have been validated using another data source, it is possible there was some degree of misclassification with samples classified as
RAS tested but in practice not tested for all exons 2, 3, 4 of
KRAS and
NRAS.
Despite efforts to obtain a higher response rate following the first two rounds of the study, the response rate for Round 3 was low (5.9%). This could potentially introduce selection bias as shown by the relatively high volume of mCRC patients treated by the participating oncologists (median of 40 in the past 3 months) [
21,
22]. For the MRR study, a similarly low response rate was observed amongst oncologists in Round 1. A higher response rate was observed in Rounds 2 and 3 after changing to a more targeted methodology without this impacting the study results [
18]. Finally, response rates of <10% are not uncommon for knowledge physician surveys [
23].
As has been described elsewhere,
RAS testing methods have been refined considerably over the last five years [
24‐
26], and the results of Round 3 of the MRR are in agreement with this, showing that nearly all (94.1%) of the pathology laboratories surveyed regarding their
RAS testing practices reported using a CE-marked or otherwise validated
RAS mutation detection method and that all had participated in at least one QA scheme. However, there is still clear need for improvement, potentially via additional education, to ensure that all oncologists and pathologists treating patients with mCRC are implementing full
RAS testing.
Competing interests
JHvK has received honoraria and research grants from AMGEN and Merck Serono.
GK is a compensated employee of Amgen Ltd. as an Observational Research Senior Manager and a stockholder in Amgen Ltd.
JB provided statistical input into the design and analysis of the studies, as a contract Biostatistician funded by Amgen Ltd., and has no competing interests to declare.
LM has provided a consulting or advisory role and received honoraria, travel, accommodation and expenses from Sanofi, Amgen, Baxter, Lilly and Roche, and has patents, royalties or other intellectual property with Convergance.
JTo has provided a consulting or advisory role and received honoraria, travel, accommodation and expenses from Amgen, Bayer, Lilly, Roche and Sanofi.
ER has received funding for consultation and presentations from AstraZeneca and Roche.
PF has participated in advisory boards for Amgen and Pfizer and has received honoraria as a speaker for Amgen Ltd., Merck Serono, Astra Zeneca, F. Hoffmann-La Roche and Pfizer.
GDM has no competing interests to declare.
PGA has participated in advisory boards and has received honoraria as a speaker for Amgen, Merck Serono, F. Hoffmann-La Roche, Sanofi, Bayer and Pfizer.
GA has received educational grants from Roche, Eli-Lilly, Amgen, Merck, Bayer and Sanofi in the last two years.
PP is a compensated employee of Amgen Ltd. as a Clinical Research Study Manager.
GDo is a compensated employee of Amgen Ltd. as a Biostatistics Senior Manager and a stockholder in Amgen Ltd.
GDe is a compensated employee of Amgen Ltd. as an Oncology Medical Team Lead and a stockholder in Amgen Ltd.
JTr has participated in advisory boards and received honoraria, travel, accommodation and expenses from Amgen, Merck Serono, Roche and Bayer.