Introduction
Colorectal cancer (CRC) is the third most frequently diagnosed cancer worldwide and the second leading cause of cancer-associated deaths. Unfortunately, 20% of patients show distant metastases at diagnosis and about 35% will develop metastases during the course of the disease (1). In 75-90% of advanced cases, it is not possible to resect all lesions and thus treatment with palliative chemo-immunotherapy is applied (2). Although current first- and second-line therapies consider genetic characteristics, like KRAS/NRAS/BRAF mutations, microsatellite instability (MSI), and other factors, including tumour location, age, and co-morbidities (3), the therapeutic mainstay consists of 5-Fluorouracil (5-FU)-based chemotherapy.
Generally, 5-FU or its prodrug capecitabine is applied with or without irinotecan and/or oxaliplatin, plus targeted immunotherapy, like the anti-vascular endothelial growth factor (VEGF) antibody bevacizumab or the anti-epidermal growth factor receptor (EGFR) agents cetuximab or panitumumab (4). Anti-EGFR antibodies are less effective in right-sided CRC and not effective in RAS-mutated CRC (5,6).
5-FU is an antimetabolite, which, due to its structural similarity to the ribonucleic acid uracil, gets incorporated into RNA instead of uracil and leads to the inhibition of DNA biosynthesis and cell growth. It also inhibits the thymidylate synthase, an enzyme important for pyrimidine synthesis. 5-FU is often combined with Leucovorin® (folinic acid), a chemo-protectant that blocks the side effects of 5-FU and is used for the potentiation of the 5-FU tumouricidal effects. It increases patients' survival and response rates (7). Oxaliplatin is a platinum compound and causes inhibition of DNA synthesis by crosslinking DNA (8). Irinotecan is an analogue of camptothecin, which is converted to its active metabolite in the body. Irinotecan inhibits topoisomerase 1, causes inhibition of DNA synthesis and DNA double strand breaks, and ultimately leads to cell cycle arrest and cell death (9).
The combination-chemotherapy regimens FOLFOX (folinic acid, 5-FU and oxaliplatin; CAPOX when oral capecitabine is applied) and FOLFIRI (folinic acid, 5-FU and irinotecan; CAPIRI when oral capecitabine is applied) show equal efficacy, however, have a different toxicity profile, which can guide decision-making (10).
The triplet FOLFOXFIRI can be used in certain cases, e.g. for fit patients without co-morbidities or for patients harbouring a BRAF mutation (10).
In mCRC patients, current second-line treatment depends mainly on the prior first-line treatment. Patients, who have received an oxaliplatin-containing doublet combination therapy, are switched to an irinotecan-containing doublet, and vice versa. Again, in second-line treatment, anti-EGFR monoclonal antibodies (mAbs) and bevacizumab can be added to the chemotherapy. Aflibercept is an alternative anti-VEGF agent, which can be applied in second-line therapy (10). Anti-VEGF agents can be effective, although the patient progressed during first-line treatment. Until recently, anti-EGFR antibodies were considered only useful in one line of treatment. New treatment options suggest that reintroduction or continuing use of anti-EGFR therapy can be beneficial, if no KRAS mutations have occurred during the initial treatment (7,11).
However, tumours can be resistant against chemotherapeutic agents, either prior to treatment start (‘innate chemoresistance’) or during the course of treatment (‘acquired chemoresistance’). Often tumours are not only resistant to one drug, but also to other or similar agents (‘multidrug-resistance’) (7,12). Many mechanisms, such as elevated metabolism, enhanced drug efflux, increased DNA repair capacity, growth factors, or genetic and epigenetic factors are involved in the process of chemoresistance, but this process is yet not fully understood (13).
This study thus aimed to investigate the role of multidrug-resistance in a real-world setting: this study explored, if under current first- and second-line protocols for mCRC, response to first-line therapy can predict response to second-line therapy. The null-hypothesis of this study is that the rates of remissions in second-line chemo-immunotherapy do not differ between patients with CRC stage IV, who had a response in first-line chemo-immunotherapy, and patients, who had no response in first-line chemo-immunotherapy.
Discussion
Cancer research has aimed for decades to better understand carcinogenesis, tumour biology and host-tumour interactions in order to find more specific and less toxic therapies. For mCRC, several important biological factors such as tumour sidedness, KRAS/NRAS/BRAF mutations, or microsatellite instability have been described and therapy algorithms were developed in order to optimally treat these patient subgroups.
Nonetheless, for the majority of patients, first- and second line-therapy of mCRC is usually based on combination chemotherapy with a 5-FU backbone, if not for BRAF-mutated or MSI-high tumours. Thus, therapy-resistance could be an important factor, based on the high similarity of first- and second-line treatment. We therefore investigated this factor in a retrospective cohort study and could clearly demonstrate that patients, who do not respond to first-line treatment have little benefit of second-line treatment. In our cohort none of the patients, who had PD to first-line therapy responded to second-line therapy (ORR=0%). This dramatic finding has to be further investigated in larger studies and the true percentage of ORR is for sure not zero; however, it clearly demonstrates the limited benefit of current treatment strategies for these patients and the high need for different therapy approaches.
Unfortunately, clinical data on response to second-line treatment in mCRC patients not responding to first-line treatment is scarce. To the best of our knowledge, we could not find one single randomized clinical trial (RCT) nor prospective analysis addressing specifically this question. There are numerous RCTs evaluating in general the efficacy and toxicity of second-line systemic therapy in mCRC patients, where disease progressed, recurred or did not respond to first-line systemic therapy.
Due to the retrospective nature of this study, mechanistic aspects of therapy-resistance could unfortunately not be investigated, but it seems most likely, that innate chemoresistance plays a major role. Biological factors for chemo- or multidrug-resistance comprise numerous mechanisms, such as elevated metabolism, enhanced drug efflux, increased DNA repair capacity, growth factors, or genetic and epigenetic factors (13,17,18). Antimetabolites, such as 5-FU cause base lesions, which promote replication fork stalling in proliferating cancer cells. Here, resistance can be acquired by e.g. stabilization of these stalled replication forks (19). Another mechanism for 5-FU resistance is enhancement of poly [ADP-ribose] polymerase 1 (PARP1) activity, increasing the base excision repair capacity in cancer cells (20). As PARP-inhibitors are broadly applied in clinical oncology with a favourable toxicity profile, these substances could be interesting candidates for combination strategies in order to overcome chemo-resistance.
For oxaliplatin, which induces intra-strand dinucleotide DNA adducts that have to be repaired again by nucleotide excision repair mechanisms, it could be shown, that upregulation of the high-mobility group A 2 gene (HMGA2) could induce oxaliplatin-resistance (21).
Eventually, for irinotecan, it could be shown, that also its mode of action (trapping the topoisomerase I, leading to replication fork stalling and collapse and cytotoxic double strand breaks), could be counterfeit by HGMA2, as HGMA2 can inhibit topoisomerase I trapping (17,22).
This preclinical evidence of cellular mechanisms, especially of HGMA2 conveying resistance to two important drugs for chemotherapy of mCRC cells, should underline the high importance of multidrug-resistance.
To date, clinical research data on chemo-resistance in CRC is unfortunately still scarce.
Due to the retrospective nature of this study, also our data has numerous limitations. As mentioned before, in the first years of the observation period, molecular biologic testing for BRAF and MSI was not routinely established and therefore especially data on these patient groups have many missing variables. Thus, the OS curves in Fig. 3 stratify only for patient or tumour characteristics but not in conjunction with first- or second-line therapy responses in these respective subgroups, as these subgroups are too small for robust interpretation. In future studies, special emphasis should be put on these patient groups in order to evaluate the effect of stratified therapies compared to 5-FU based second-line chemo-immunotherapy.
Moreover, retrospective studies always harbour the risk of selection bias and the influence of confounding variables.
In the future, patients with PD to first-line therapy should be tested thoroughly by gaining novel tumour specimens, which have to be comprehensively tested, including for known factors of multidrug-resistance. Ideally, detected alterations should be compared to the original specimen harboured during primary diagnosis in order to differentiate between acquired and innate chemoresistance. These patients should be regarded as ‘functional high risk’ and thus primarily treated within clinical trials. Based on the comprehensive testing results, targeted therapies or immunotherapies should be preferred to conventional chemotherapies. If the comprehensive profiling lacks specific targets or biomarkers for response to immunotherapy, clinical trials that specifically address chemo-resistance should be performed in this patient group.
For another proportion of these patients, especially frail ones and patients that suffered from severe toxicities of first-line therapy, BSC might be an option in order to avoid further side effects of futile therapies. So far, the current National Comprehensive Cancer Network (NCCN)-guidelines recommend BSC as second line option for patients with PD after first line and no improvement to functional status, or after failure to second-line treatment (23).
On the long run, however, novel and different therapy approaches have to be developed in order to benefit also this precarious patient group. This is most effectively done in randomized clinical trials and thus future studies should focus especially on this patient group, because of their high medical need.
Besides its retrospective nature, one big limitation of this study is its small sample size due to the design as single-centre study. Moreover, not all patients could be included, because they had to receive at least two lines of chemo-immunotherapy to evaluate PFS1, PFS2 and OS. Furthermore, not all patients, which were included in the study, were tested for all genetic markers, thus leading to missing data. Because of the small sample size, also subgroup analyses could not be performed. Future studies should be large enough to do subgroup analyses in order to rule out confounding factors such as co-morbidities, treatment response time, and other clinical characteristics.
In conclusion, this study demonstrates that, with current treatment strategies applying 5-FU-based chemo-immunotherapy in first-as well as second-line treatment of mCRC patients, response to first-line therapy is a strong predictor for response in second-line and OS. By only exchanging the chemotherapeutic combination partner and additive antibody, the negative factor of not responding to first-line therapy, most likely caused by multidrug-resistance, could not be overcome in this study population.
These findings have to be confirmed in larger studies, but raise the need for more basic research in CRC and on multidrug-resistance in order to gain novel treatment options, especially for patients not responding to first-line 5-FU-based chemo-immunotherapy.