Inclusion of targeted therapies in the standard of care for metastatic colorectal cancer patients in a German cancer center: the more the better?!

Abstract

Purpose

Significant prolongation of overall survival (OS) has been reached in metastatic colorectal cancer (mCRC) treatment within the last 5–10 years. Our study was conducted in order to evaluate and compare OS of different standard of care treatment options in a university-based outpatient clinic.

Methods

One hundred and three mCRC patients were identified by retrospective analysis and treated according to available guidelines. OS was analyzed according to the different combinations of first- and second-line treatments.

Results

mCRC patients revealed an mOS of 34.4 months. Patients receiving anti-vascular endothelial growth factor (VEGF) blockade in at least one treatment line showed a significantly longer survival time (p = 0.0056) versus patients without any bevacizumab. No OS differences were detected comparing the different first- and second-line chemotherapy (CTX) strategies in the unselected population. However, wild-type (wt) Kras patients treated with anti-epidermal growth factor receptor (EGFR) therapy plus CTX in first-line therapy showed significantly longer OS compared to those receiving only additional VEGF inhibition or no targeted therapy (p = 0.0056; mOS 46.8 vs. 20.4 months, respectively). wt Kras patients profited in trend (p = 0.076) from CTX combinations of first-line anti-EGFR followed by second-line anti-VEGF compared to first-line anti-VEGF followed by second-line anti-EGFR (mOS 46.8 vs. 19.2 months, respectively).

Conclusions

Our results indicate successful allocation of the current mCRC treatment according to the Kras status. Differences in OS of wt Kras patients indicated the further need for randomized trials to define the potential benefit of sequential therapy with EGFR inhibition in first-line therapy followed by VEGFR inhibition vice versa.

Background

Colorectal cancer (CRC) is the third most common malignant neoplasm worldwide (Shike et al. 1990) and the fourth most common cause of cancer-related deaths (International Agency for Research on Cancer (2010) GLOBOCAN 2008 version 2.0, Estimated cancer incidence, mortality prevalence, and disability-adjusted life years (DALYs) worldwide in 2008). Based on the data collected in the SEER database for US patients, median overall survival (OS) for metastatic CRC (mCRC) improved from 8 months in 1988 to 14 months in 2007 (Lawrence et al. 2012). Considerable efforts and intensive clinical research over the last 5–10 years focused on the efficiency of targeted inhibition in addition to traditional treatment strategies. Recently, the CRYSTAL study (Van Cutsem et al. 2009) and PRIME study (Douillard et al. 2010) demonstrated that first-line cetuximab and panitumumab in chemotherapy (CTX) combinations resulted in a 3.5 months (23.5 vs. 20 months, respectively; p = 0.0094) and 4.2 months (23.9 vs. 19.7 months, respectively; p = 0.072) improvement of OS compared to FOLFIRI or FOLFOX4 treatment alone. Similarly, bevacizumab in first-line treatment of mCRC resulted in a comparable advantage in median OS of 3.2 months when tested on an irinotecan, bolus fluorouracil (FU), and leucovorin backbone (18.3 and 15.1 months, respectively (Hurwitz et al. 2005)). Efficacy analysis of bevacizumab on an oxaliplatin-based CTX backbone showed only modest improvement of median OS (21.3 vs. 19.9 months, respectively; p = 0.077; (Saltz et al. 2008)). Notably, a direct comparison of epidermal growth factor receptor (EGFR) versus vascular endothelial growth factor (VEGF) inhibition strategies in first line based on the mentioned studies was potentially erroneous due to different CTX backbones administered. However, preliminary data from a clinical phase III trial (Heinemann et al. 2013) comparing the overall response rates of the first-line strategies of bevacizumab versus cetuximab treatment on the backbone of a FOLFIRI CTX became available recently. Results indicate that the use of EGFR inhibition in patients with a Kras wild-type (wt) results in a slight benefit in OS.

Even less data were available to compare EGFR and VEGF inhibition in the second-line treatment [for review, see (Chibaudel et al. 2012)]. Isolated investigations of second-line treatments remain biased with respect to OS since patients in control arms frequently received the study drug in third-line or subsequent treatments (Lenz et al. 2006; Van Cutsem et al. 2007). Continuation of bevacizumab treatment beyond first-line progression of disease proved to be advantageous over discontinuation of VEGF inhibition (OS 11.2 vs. 9.8 months, p = 0.0062 (Bennouna et al. 2013)). No comparable data were available for continuation of EGFR inhibition.

Together, recent data from phase III clinical trials showed that the addition of targeted biological agents to first- or second-line treatment could significantly improve disease outcome [for review, see (Chibaudel et al. 2012)]. Since both regimens were mechanistically different and potentially complementary, it was of enormous practical relevance whether patients might benefit from a second-line crossover strategy of EGFR/VEGF inhibition rather than a continuation of treatment. Further, comparative analyses of survival data from clinical trials and statistical data from high-volume centers indicated substantial differences between rigorously controlled (i.e., in- and exclusion criteria) clinical trials and standard of care situation (i.e., real-life experiences) in typical outpatient clinics. In order to evaluate practical outcome of different treatment regimens in an outpatient clinic, we have therefore analyzed 112 patients treated for mCRC at the University Medical Centre Mainz, Germany, between 2001 and 2012, primarily not included into clinical trials. Hereby, primary focus of the analysis was patient outcome as determined by OS observed in different combinations of anti-VEGF and anti-EGFR antibody approaches.

Patients and methods

Patient population

The unselected and consecutively treated patient population with advanced colorectal cancer between January 1, 2001, and December 31, 2011, were identified in the outpatient clinic of gastrointestinal oncology of the University Medical Centre Mainz, Germany. Data cutoff was December 31, 2012. Patients had to be over 18 years old and gave written informed consent for scientific analysis of their data.

Treatment strategies

Patients received treatment in line mostly within approved labels and the German S3-guidelines. First-line treatment included a backbone of an irinotecan-fluoropyrimidin or oxaliplatin–fluoropyrimidin-based CTX with the addition of a VEGF (bevacizumab), EGFR (cetuximab, panitumumab), or RTK (sunitinib, sorafenib, imatinib, gefitinib) inhibitor when appropriate, as described before (Awada et al. 2011; Blesa and Pulido 2010; Cutsem et al. 2007; Fisher et al. 2008; Hurwitz et al. 2004; Kelley et al. 2013; Kindler et al. 2005; Peeters et al. 2010; Qvortrup et al. 2010; Samalin et al. 2014; Tournigand et al. 2004). As a second-line treatment backbone CTX was continued in a crossover design, while VEGF, EGFR, or RTK inhibition was either discontinued or switched over (see Fig. 1a–c for overview). Criteria for discontinuation or change of the applied regimes were progressive disease (PD), intolerable toxicities defined by CTC > grade 3, or refusal of further treatment. First-or second-line treatment was additionally stopped in case of resectability. In case of third- and fourth-line protocols, 5-FU monotherapy or mitomycin c with or without capecitabine were applied.

Fig. 1.

First- and second-line treatment combinations. a Second-line continuation—targeted therapy was continued beyond progression. b Second-line crossover—targeted therapy was changed: anti-VEGF first-line treatment was switched to anti-EGFR second-line; anti-EGFR first-line therapy was switched to anti-VEGF second line. c Second-line optional—no targeted therapy in first line, second line, or both. First-line CTX: irinotecan–fluoropyrimidin or oxaliplatin–fluoropyrimidin based; second-line CTX crossover; anti-VEGFR: bevacizumab treatment; anti-EGFR: cetuximab or panitumumab

Safety and efficacy assessment

Undesirable side effects (USE) were recorded for all treatment effective cases. Evaluation occurred with respect to the reason for end of the treatment.

Tumor response was routinely measured every 8–10 weeks by CT scan and assessed and graded based on RECIST criteria, version 1.1. Patients were followed up to December 31, 2012, (censored) or until death occurred.

Endpoint of analysis

OS consisted of the primary endpoint of the study and was defined as time from date of diagnosis to death. Additional endpoints were time to failure of strategy (TTFS) (defined as start of a new treatment line), DCR (sum of complete responses (CR) + partial responses (PR) + stable disease (SD) + mixed response), best response as well as resectability as a result of the different first-line strategies.

Kras analysis

For each tumor sample, DNA was extracted from two 5 µM formalin-fixed paraffin embedded FFPE tumor sections (deparaffinized), using the High-Pure PCR Template Preparation Kit (Roche Diagnostic Spa, Indianapolis, USA), according to the manufacturer’s instructions. DNA (25 ng) was amplified using oligonucleotide primers specific for human KRAS (exons 12 and 13) (Moehler et al. 2010). PCR reactions were run on a LightCycler 480 (LC480) with the High Resolution Melting Master (Roche Diagnostic Spa, Indianapolis, USA), according to the manufacturer’s instructions. The data were analyzed with the LightCycler 480 Gene Scanning Software Module for deletion and mutation identification.

Statistical analysis

One hundred and twelve patients with histologically proven mCRC were initially selected from the data base. Date of diagnosis and date of death (or censoring date) could be verified for 103 patients (primary analysis population). OS and TTFS were evaluated by Kaplan–Meier estimates. The analysis of additional endpoints and all further data were interpreted descriptively.

Results

Patient characteristics

Demographic and baseline characteristics are provided in Table 1. Among 112 patients documented, 103 patients (31 females; 72 males) fulfilled criteria for inclusion (date of first diagnosis, date for start of first line and date of death/censoring date). At date of diagnosis, 33 % of patients were below 60 years, 40 % between 60 and 70 years, and 27 % above 70 years of age (median 63 years). Tumor localization was equally distributed between rectum (n = 34, 33 %) and sigma (n = 32, 31 %) and the rest of the colon (n = 30, 29 %), while only 7 % of patients presented with carcinoma of the coecum (n = 7). The majority of patients presented with an initial T3 tumor stage (60 %) and N2 nodal status (48 %). In 89 % of patients, liver metastases were diagnosed while 47 % presented with metastases in 3 or more organs. Kras status was determined in 71 patients, 79 % (n = 56) of which were Kras wt, while Kras mutation was detected in 21 % (n = 15) of patients.

Table 1.

Baseline characteristics

Baseline characteristics	N	%
Number of patients	103	100.0 %
Age (years)
<60 years	34	33
≥60 to ≤70 years	41	40
≥70 years	28	27
Gender
Male	72	70
Female	31	30
Location of primary tumor
Colon	30	29
Coecum	7	7
Rectum	34	33
Sigma	32	31
Metastases: localization
Liver	92	89
Lung	52	51
Bones	14	14
Peritoneum	37	36
Lymph nodes	44	43
other	20	19
Metastases: number of organs
≥3 organs	48	47
<3 organs	55	53
Kras status
wt	56	54
Mutation	15	15
NK	32	31
Number of treatment lines started
1	22	21
2	33	32
3	19	18
4	12	12
≥5	17	17

Treatment

Twenty-one percentage (n = 22) of patients received first-line treatment only, while 33 % (n = 34), 18 % (n = 19), and 12 % (n = 12) received a second-, third- or fourth-line therapy, respectively. With respect to a targeted therapy, 64 % (n = 66) received bevacizumab, while 62 % (n = 64) received anti-EGFR therapy (cetuximab/panitumumab) in one or more treatment lines (see Fig. 1 for overview). Furthermore, 23 % of patients (n = 24) were treated with tyrosine kinase inhibitors (TKIs) such as sunitinib, imatinib, gefitinib, or sorafenib in second-, third- or fourth-line therapy after inclusion into clinical trials or requests to health insurances for reimbursement. The performance status of the patients did not vary between the examined groups.

Efficacy

The mOS of all patients treated was 34.8 months. Stratification in accordance with age (<60; ≥60–70; ≥70 years), Kras status or the number of organs with metastases (<3 organs; ≥3 organs) showed no significant difference in terms of OS. Patients with coecum cancers showed a trend (p = 0.09) toward reduced mOS [19.2 months, 95 % CI (1.2–25.2)] compared to patients with colon-, rectum-, or sigma-adenocarcinoma [30 months (15.6–48); 36 months (25.2–45.6); 30 months (18; 46.8), respectively].

First-line treatment

Irrespective of treatment used, secondary resectability was reached in 10 % (n = 10) of the unselected patient population and treatment was stopped due to toxicities in 7 %, while 3 % deceased before starting the next treatment line. DCR was reached in 72 % (n = 74) of patients. No significant difference in terms of OS could be detected between patients receiving an oxaliplatin-containing regime [mOS 27.6 months, 95 % CI (18–39.6)] or an irinotecan-containing regime (mOS 33.6 months, 95 % CI (24–46.8)] as well as between patients treated with an anti-VEGF or anti-EGFR antibody in addition to the CTX (Table 2). However, patients with Kras wt profited significantly more in terms of OS under EGFR inhibition in comparison with those who received anti-VEGF treatment (Fig. 2, p = 0.0056; mOS 46.8 vs. 20.4 months, respectively). No significant differences were detected in TTFS of different first-line strategies (data not shown).

Table 2.

First-line treatment: Kaplan–Meier estimates for mOS

First-line treatment	N	%	mOS	Significance
Number of patients	103	100	Months; 95 % CI	Log rank
CTX backbone
Oxaliplatin	41	40	33.6 (24–46.8)	p = 0.24
Irinotecan	58	56	27.6 (18–39.6)
Fluoropyrimidin mono	4	4	49.2 (34.2–.)
Targeted therapy
Anti-VEGF	37	36	27.6 (18–45.6)	p = 0.79
Anti-EGFR	28	27	42 (20.4–72)
None	38	38	30 (24–44.4)
CTX combinations
Oxaliplatin/anti-EGFR	3	3	42 (–)	p = 0.47
Oxaliplatin/anti-VEGF	18	18	25.2 (16.8–48)
Oxaliplatin/none	20	19	26.4 (12–36)
Irinotecan/anti-EGFR	25	24	46.8 (15.6–72)
Irinotecan/anti-VEGF	19	18	28.8 (15.6–46.8)	p = 0.86
Irinotecan/none	14	14	33.6 (16.8–69.6)

Fig. 2.

Kaplan–Meier estimate for OS: patients with Kras wt stratified by targeted first-line therapy

Second-line treatment

Irrespective of treatment used, resectability was reached in 5 % (n = 4) of the patients. DCR was reached in 46 % (n = 37). We detected no significant difference regarding the impact of the applied second-line therapy to patients’ survival (Table 3). Furthermore, patients with Kras wt profited equally from the addition of anti-VEGF or anti-EGFR to the backbone CTX in the second line (data not shown). Among the main reasons for second-line EOT, we recorded 65 % PD, 15 % treatment interruption, 5 % reached resectability, and 6 % had toxicities.

Table 3.

Second-line treatment: Kaplan–Meier estimates for mOS

Second-line treatment
Number of patients	N 81	% 100	mOS Months; 95 % CI	Significance Log rank
CTX backbone
Oxaliplatin	37	46	n.d	n.d.
Irinotecan	40	49
Fluoropyrimidin mono	3	4
No CTX	1	1
Targeted therapy
Anti-VEGF	26	32	26.4 (12–28.8)	p = 0.76
Anti-EGFR	21	26	14.4 (4.8–24)
None	34	42	16.8 (8.4–27.6)
CTX combinations
Oxaliplatin/anti-EGFR	4	5	n.d.	n.d.
Irinotecan/anti-EGFR	17	21
Oxaliplatin/anti-VEGF	15	19
Irinotecan/anti-VEGF	10	12
Oxaliplatin/none	18	22
Irinotecan/none	13	16

First- and second-line treatment combinations

We detected no significant difference in OS regarding the order and identity of different first- and second-line treatment combinations (Table 4). However, Kras wt patients profited in trend under first-line anti-EGFR followed by second-line anti-VEGF treatment in comparison with those who received first-line anti-VEGF followed by second-line anti-EGFR treatment (p = 0.076; mOS 46.8 vs. 19.2 months, respectively). Using a 1-year landmark analysis, this difference almost reaches significance (p = 0.058), resulting in a difference of 3-year survival rates of 75 % [95 % CI (12.8–96.1) vs. 25 % 95 % CI (0.9–66.5)], respectively.

Table 4.

First- and second-line combinations: Kaplan–Meier estimates for mOS

First- and second-line combinations
Number of patients	N 81	% 100	mOS Months; 95 % CI	Significance Log rank
Targeted therapy
Crossover	13	16	21.6 (18–46.8)	p = 0.78
Treatment beyond pd	12	15	33.6 (14.4–46.8)
None in first line	22	27	34.8 (22.8–48)
None in second line	20	25	27.6 (15.6–48)
None	14	17	28.8 (13.2–54)

Overall comparison of targeted therapies

Patients receiving anti-VEGF treatment in at least one treatment line showed significant advantage compared to those not receiving anti-VEGF [Fig. 3, p = 0.0056; mOS 34.8 months, 95 % CI (27.6–46.8) vs. mOS 25.2 months, 95 % CI (13.2–30), respectively]. Interestingly, no significant benefit with regard to OS was detected when anti-EGFR or TKIs treatment was tested (p = 0.89 and p = 0.835, respectively).

Fig. 3.

Kaplan–Meier estimate for OS: patients stratified by Bevacizumab treatment in at least one treatment line

Discussion

Evaluation of new treatment options in mCRC by clinical trials remains a major challenge due to the selection of appropriate endpoints and elimination of bias. However, final proof of a patient benefit is survival time in a standard of care situation, when treatment decisions have to be made on a case-to-case basis. Stratification of first-line CTX backbones oxaliplatin and irinotecan (irrespective of targeted biological agents added) in our study population revealed no significant difference with respect to OS. The median OS rate of 30–36 months for unselected patients successfully validated the selection of treatment regimens evaluated by recent large phase III clinical trials. Most recent results from FIRE III (Heinemann et al. 2013) and TRIBE trial (Loupakis et al. 2013), comparing first-line treatments of mCRC, showed a median OS for FOLFIRI + cetuximab and FOLFOXIRI + bevacizumab of 28.7 and 31 months, respectively. Comparing our findings with results from previous trials without targeted biological agents (e.g., (Falcone et al. 2007), mOS for the beneficial treatment 22.6 months) indicated a difference in mOS of roughly 10 months. Taking into account that detection of disease progression created a psychologically unfavorable situation which might influence the course of disease, it is worth mentioning that a DCR of 72 % was reached and the TTFS was quite comparable between the different first-line strategies.

In the FIRE III (Heinemann et al. 2013) trial, the first-line treatment options of anti-VEGF and anti-EGFR therapies of mCRC were compared directly for the first time. Even if the trial for its primary endpoint of objective response rate was negative, the results show an advantage in mOS of 2.7 months for first-line cetuximab versus bevacizumab on FOLFIRI CTX backbone in 592 wt Kras mCRC patients. Our mCRC population comparison of anti-VEGF and anti-EGFR treatment showed improved mOS for anti-EGFR treatment (27.6 vs. 42 months) which reached significance in the wt Kras patient subgroups. This large difference might be due to the imbalance and low numbers of patients with coecum cancers in the anti-VEGF compared to the anti-EGFR population (n = 8 vs. n = 0, respectively) which showed a trend (p = 0.09) toward reduced median OS (19.2 months) compared to the other indications (≥30 months). Taking into consideration the translational results of the FIRE III and PRIME study (Oliner et al. 2013; Stintzing et al. 2013), it is tempting to assume that the wt Kras patients subgroups in this trial might also be negative for other Ras mutations, which were not determined though in this study. This might provide a possible explanation for the distinct difference in OS in favor for anti-EGFR treatment (Fig. 2). However, our outcome is in line with results from the FIRE III trial. Similar to the analysis of PFS in the FIRE III trial, time to failure of treatment strategy hardly showed any difference in our population between the subgroups analyzed.

Analysis of OS suggested an advantage for a second-line treatment strategy containing anti-VEGF in mCRC. However, the difference in median OS of 26.4 versus 14.4 months for any anti-VEGF therapy lacked statistical significance and wt Kras patients hardly showed any difference at all. Comparing our respective outcome with phase III studies for second-line treatment revealed a quite good performance of our standard of care at the cancer center. Median OS in the ML 18147 trial (Bennouna et al. 2013) which investigated continuation of bevacizumab after first progression in mCRC was 11.2 months for bevacizumab plus CTX and 9.8 months for CTX alone (p = 0.0062). The discrepancy in the mOS between our results and the ML 18147 trial can be attributed to the heterogeneity of the study design since no patients with anti-EGFR agents in the first-line treatment were included in the ML 18147 trial. Furthermore, our study is a retrospective analysis of data and deviations in direct comparison with prospective trials are expected. The respective EPIC phase III trial for evaluation of anti-EGFR inhibition in second-line treatment of wt mCRC patients revealed mOS of 10.7 months with cetuximab/irinotecan and 10.0 months with irinotecan alone without statistical significance (Sobrero et al. 2008). For our Kras wt population, our standard of care treatment resulted in mOS of 16.8 and 12 months for CTX/anti-EGFR and CTX alone, respectively. In line with the findings of the EPIC trial, we observed no statistical significance between the two groups.

Our analysis of the different combinations of first- and second-line therapies stratified as crossover versus continuation of targeted therapy revealed no statistically significant differences. However, the population receiving a crossover treatment reached quite high OS values. Subgroup analyses for a preferred treatment order showed a trend toward first-line anti-EGFR followed by second-line anti-VEGF treatment. This population contained a group of patients (N/n 13/7) with undetermined Kras status which received anti-EGFR treatment in second line. Comparing these two strategies in the small subgroup of wt Kras patients was more conclusive. An 1-year landmark analysis showed a difference in 3-year survival of 75 versus 25 % which might be worth analyzing in a prospective randomized clinical trial.

On the basis of these results, it is remarkable that patients receiving anti-VEGF treatment in at least one treatment line clearly showed an advantage in OS indicating that anti-VEGF is a desirable component of individualized mCRC treatment.

Conclusions

Together, the key findings of our study at a German cancer center are (1) implementation of recent developments from clinical trials investigating mCRC treatment has raised mOS for mCRC patients above 30 months in the standard of care setting; (2) anti-VEGF treatment in at least one treatment line results in an OS benefit for mCRC patients; (3) in the population of wt Kras patients, first-line EGFR inhibition followed by second-line VEGF inhibition seems to be the preferable order of treatments.

Abbreviations

OS

Overall survival

mCRC

Metastatic colorectal cancer

CTX

Chemotherapy

VEGF

Vascular endothelial growth factor

EGFR

Epidermal growth factor receptor

TK

Tyrosine kinases

TTFS

Time to failure of strategy

DCR

Disease control rate

CR

Complete response

PR

Partial response

SD

Stable disease

USE

Undesirable side effects

PD

Progressive disease

wt

Wild type

EOT

End of treatment