- Email: [email protected]
Biomarkers of Tumour Radiosensitivity and Predicting Benefit from Radiotherapy
Clinical Oncology xxx (2015) 1e9 Contents lists available at ScienceDirect
Clinical Oncology journal homepage: www.clinicaloncologyonline.net
Overview
Biomarkers of Tumour Radiosensitivity and Predicting Benefit from Radiotherapy L.J. Forker *y, A. Choudhury *y, A.E. Kiltie z * Institute
of Cancer Sciences, The University of Manchester, Manchester Academic Health Science Centre, The Christie NHS Foundation Trust, Manchester, UK y Department of Clinical Oncology, The Christie NHS Foundation Trust, Manchester, UK z CRUK/MRC Oxford Institute for Radiation Oncology, Department of Oncology, University of Oxford, Old Road Campus Research Building, Headington, Oxford, UK Received 5 February 2015; accepted 2 June 2015
Abstract Radiotherapy is an essential component of treatment for more than half of newly diagnosed cancer patients. The response to radiotherapy varies widely between individuals and although advances in technology have allowed the adaptation of radiotherapy fields to tumour anatomy, it is still not possible to tailor radiotherapy based on tumour biology. A biomarker of intrinsic radiosensitivity would be extremely valuable for individual dosing, aiding decision making between radical treatment options and avoiding toxicity of neoadjuvant or adjuvant radiotherapy in those unlikely to benefit. This systematic review summarises the current evidence for biomarkers under investigation as predictors of radiotherapy benefit. Only 10 biomarkers were identified as having been evaluated for their radiotherapy-specific predictive value in over 100 patients in a clinical setting, highlighting that despite a rich literature there were few highquality studies for inclusion. The most extensively studied radiotherapy predictive biomarkers were the radiosensitivity index and MRE11; however, neither has been evaluated in a randomised controlled trial. Although these biomarkers show promise, there is not enough evidence to justify their use in routine practice. Further validation is needed before biomarkers can fulfil their potential and predict treatment outcomes for large numbers of patients. Ó 2015 Published by Elsevier Ltd on behalf of The Royal College of Radiologists.
Key words: DNA damage response; molecular signature; predictive biomarker; radiosensitivity; radiotherapy
Statement of Search Strategies Used and Sources of Information A literature search of PubMed for molecular signatures predictive of radiotherapy benefit was conducted using the terms (predict OR prediction OR predictive OR predictor OR predicts) AND (radiotherapy OR radiosensitivity OR chemoradiotherapy) AND (‘radiosensitivity index’ OR gene signature OR molecular signature OR gene expression profile) and yielded 122 results. A second literature search of PubMed for biomarkers related to DNA damage response predictive of radiotherapy benefit using the terms (predict
Author for correspondence: A. Choudhury, Institute of Cancer Sciences, The University of Manchester, Manchester Academic Health Science Centre, The Christie NHS Foundation Trust, Wilmslow Road, Manchester M20 4BX, UK. Tel: þ44-161-446-3000; Fax: þ44-161-446-3977. E-mail address: [email protected] (A. Choudhury).
OR prediction OR predictive OR predictor OR predicts) AND (radiotherapy OR radiosensitivity OR chemoradiotherapy) AND (biomarker OR molecular OR signature) AND (DNA repair OR DNA damage OR DNA OR ‘DNA damage response’) found 476 results. Searches were supplemented by hand searching of reference lists of relevant studies and reviews. Abstracts were reviewed for all relevant titles and full papers obtained if necessary. Citations were excluded if they were not original research articles or not relevant to cancer biomarker research. Biomarkers predicting benefit from chemotherapy, hypoxia modifying therapy, targeted agents and endocrine therapy were excluded, as were those concerning diagnostic or prognostic markers and those predictive of normal tissue toxicity. Citations relating to imaging markers were excluded. Pre-clinical work, studies with small patient numbers (<100) and studies with no control population were also excluded. Only reports published in English up to January 2015 week 1 were included.
http://dx.doi.org/10.1016/j.clon.2015.06.002 0936-6555/Ó 2015 Published by Elsevier Ltd on behalf of The Royal College of Radiologists.
Please cite this article in press as: Forker LJ, et al., Biomarkers of Tumour Radiosensitivity and Predicting Benefit from Radiotherapy, Clinical Oncology (2015), http://dx.doi.org/10.1016/j.clon.2015.06.002
2
L.J. Forker et al. / Clinical Oncology xxx (2015) 1e9
Introduction Radiotherapy in Cancer Treatment In 2012 there were 14.1 million new cancer diagnoses worldwide, the most common being lung (13%), breast (11.9%), colorectal (9.7%) and prostate (7.9%) [1]. Radiotherapy is recommended as a key component of curative management of these cancers, as either neoadjuvant, definitive or adjuvant treatment [2e5] and is essential in palliative oncology care [6]. Overall more than half of newly diagnosed cancer patients will require radiotherapy [7]. Although recent technological advances in radiotherapy planning and delivery techniques have led to more individualised radiotherapy based on the anatomy of the tumour with subsequent improved therapeutic ratio [8], the adaptation of radiotherapy with respect to tumour biology has significant potential to contribute to further therapeutic gain. Predictive Cancer Biomarkers Cancer is characterised by considerable genetic and epigenetic heterogeneity driven by genomic instability [9,10]. It is therefore not surprising that even patients diagnosed with the same cancer type vary widely in the natural progression of their disease and response to treatment. An era of personalised cancer medicine in which biomarkers can be used to tailor treatment to each specific patient is a major goal in oncology. Prognostic biomarkers provide information regarding disease outcome regardless of the treatment received and predictive biomarkers determine which patients will probably derive benefit from a specific therapy [11,12]. Examples of the few predictive biomarkers that have successfully transitioned to routine clinical use are those predicting response to targeted drugs, such as HER-2 in selecting breast cancer patients for HER-2 targeted therapies [13e15] and EGFR and ALK mutations in predicting response to tyrosine kinase inhibitors in non-small cell lung cancer [16e19]. These biomarkers have had a dramatic effect on current practice for selected populations of cancer patients. Radiosensitivity Biomarkers Potential applications of a radiosensitivity biomarker include tailoring of the dose depending on tumour biology, aiding decision making between radical radiotherapy and radical surgery where both are viable options and avoiding delay of definitive surgery for those unlikely to benefit from neoadjuvant radiotherapy or chemoradiotherapy. Radiosensitivity biomarkers may also be useful in predicting normal tissue toxicity, but this will be discussed in another review. Many factors are known to influence tumour response to irradiation, including total dose, fractionation, tumour potential doubling time, hypoxia and intrinsic radiosensitivity.
Fractionation and hypoxia will be discussed in separate reviews. Traditional laboratory tests to determine intrinsic radiosensitivity, such as measuring ex vivo surviving fraction at 2 Gy (SF2) by clonogenic survival assay, have been correlated with clinical outcome [20], but are not practical for routine use. Therefore alternative strategies must be sought. The development of high throughput molecular profiling techniques has led to the identification of gene expression signatures in tumours. Signatures such as Oncotype DxÒ and MammaPrintÒ have been successful in stratification for recurrence risk and aiding decision making regarding adjuvant chemotherapy in breast cancer [21,22] and a hypoxia signature has been shown to predict benefit from concurrent hypoxia modification with radiotherapy in laryngeal cancer [23]. This may also be an attractive approach for the development of a biomarker of intrinsic radiosensitivity. Additionally molecules involved in DNA damage response (DDR) signalling pathways (Figure 1), which sense and repair DNA damage [24], are excellent candidates for evaluation as radiosensitivity biomarkers, as cells with a defective DDR have less ability to repair lethal radiation-induced DNA double-strand breaks (DSBs) and are therefore more sensitive to irradiation, as seen in radiosensitivity syndromes such as ataxia telangiectasia [25]. Aims of Review Cancer biomarker research is an ever-expanding field with large numbers of publications, including numerous pre-clinical studies. In order to focus on the most promising radiosensitivity biomarkers, this review will concentrate on molecular signatures and DDR-related biomarkers that have been evaluated for their ability to predict benefit from radiotherapy in a clinical setting.
Materials and Methods Literature Searches A literature search of PubMed for molecular signatures predictive of radiotherapy benefit was conducted using the terms (predict OR prediction OR predictive OR predictor OR predicts) AND (radiotherapy OR radiosensitivity OR chemoradiotherapy) AND (‘radiosensitivity index’ OR gene signature OR molecular signature OR gene expression profile) and yielded 122 results. A second literature search of PubMed for biomarkers related to DDR predictive of radiotherapy benefit using the terms (predict OR prediction OR predictive OR predictor OR predicts) AND (radiotherapy OR radiosensitivity OR chemoradiotherapy) AND (biomarker OR molecular OR signature) AND (DNA repair OR DNA damage OR DNA OR ‘DNA damage response’) found 476 results. Searches were supplemented by hand searching of reference lists of relevant studies and reviews. Abstracts were reviewed for all relevant titles and full papers obtained if necessary.
Please cite this article in press as: Forker LJ, et al., Biomarkers of Tumour Radiosensitivity and Predicting Benefit from Radiotherapy, Clinical Oncology (2015), http://dx.doi.org/10.1016/j.clon.2015.06.002
L.J. Forker et al. / Clinical Oncology xxx (2015) 1e9
3
Fig 1. Response to DNA double-strand breaks (DSBs) induced by ionising radiation. The MRN complex (MRE11, RAD50 and NBS1) and Ku70/ Ku80 act as sensors of DSBs and activate ATM and DNA-PKcs, respectively. These phosphorylate the histone H2AX, amplifying the signal and recruiting molecules involved in the DNA damage response (DDR) to form irradiation-induced foci (IRIF). ATM phosphorylates p53 and MDM2 (a protein that normally binds p53 tagging it for destruction) causing dissociation of p53 from MDM2. p53 acts as a transcription factor, which upregulates pro-apoptotic genes and p21 (a cyclin-dependent kinase inhibitor), leading to apoptosis or G1 phase arrest, respectively. ATM also phosphorylates CHK2 to induce s phase arrest. Homologous recombination and non-homologous end joining are two mechanisms of DSB repair. In homologous recombination, MRE11 is involved in end resection to create single-stranded regions before strand invasion of homologous DNA mediated by RAD51 and XRCC2/3, which is used as a template for repair. BRCA 1 and 2 have multiple roles in homologous recombination. In nonhomologous end joining, Ku70 and Ku80 bind broken DNA ends and recruit DNA-PKcs, which phosphorylates other involved proteins. ARTEMIS processes the ends prior to ligation by LIG4, which is aided by XRCC4 [24].
Exclusion Criteria Citations were excluded if they were not original research articles or not relevant to cancer biomarker research. Biomarkers predicting benefit from chemotherapy, hypoxia modifying therapy, targeted agents and endocrine therapy were excluded, as were those concerning diagnostic or prognostic markers and those predictive of normal tissue toxicity. Citations relating to imaging markers were excluded. Pre-clinical work, studies with small patient numbers (<100) and studies with no control population were also excluded. Only reports published in English up to January 2015 week 1 were included. Figures 2 and 3 show flow diagrams of citations identified and reasons for exclusion.
Results Molecular Signatures Radiosensitivity index Torres-Roca et al. [26] were the first to show that a gene expression profile could be used to predict SF2 in cancer cell lines of numerous types [26]. A 10 gene radiosensitivity signature was developed in 48 cancer cell lines from the National Cancer Institute 60 (NCI-60) panel. Radiosensitivity was modelled as cell survival post-irradiation as a function of expression of these genes to give the
radiosensitivity index (RSI), where a lower RSI indicates higher radiosensitivity [27]. RSI was initially evaluated clinically in 14 rectal cancer patients treated with neoadjuvant chemoradiotherapy and 12 oesophageal cancer patients treated with preoperative concurrent chemoradiotherapy followed by planned esophagectomy. RSI was lower for patients who responded (defined as decrease in T stage or clinical complete response). In 92 head and neck cancer patients treated with radiotherapy and concurrent cisplatin-based chemotherapy patients with lower RSI had superior locoregional control at 2 years (locoregional control 86% radiosensitive versus 61% radioresistant, P ¼ 0.05) [27]. RSI has also been evaluated retrospectively in breast cancer in two independent cohorts. In the first cohort (n ¼ 159), radiotherapy treated patients classified as radiosensitive by RSI had improved recurrence-free survival (RFS) at 5 years (RFS 95% radiosensitive versus 75% radioresistant, P ¼ 0.0212), but there was no difference in RFS between radiosensitive and radioresistant patients in those not treated with radiotherapy (RFS 71% radiosensitive versus 77% radioresistant, P ¼ 0.6744). Most relapses were distant metastases. In a second more homogeneous cohort of 344 patients who were lymph node negative and received no adjuvant systemic treatment, radiosensitive patients had superior distant metastasis-free survival (DMFS) in the radiotherapy treated group (DMFS 77% radiosensitive versus 64% radioresistant, P ¼ 0.0409), but DMFS was similar for radiosensitive and radioresistant
Please cite this article in press as: Forker LJ, et al., Biomarkers of Tumour Radiosensitivity and Predicting Benefit from Radiotherapy, Clinical Oncology (2015), http://dx.doi.org/10.1016/j.clon.2015.06.002
4
L.J. Forker et al. / Clinical Oncology xxx (2015) 1e9
Fig 2. Flow diagram for the literature search for molecular signature biomarkers predictive of radiotherapy benefit. In total, 122 citations were identified and 117 of these were excluded (exclusion criteria are listed in the methods section). Five studies were included in the review.
Fig 3. Flow diagram for the literature search for DNA damage response-related biomarkers predictive of radiotherapy benefit. In total, 476 citations were identified and 469 of these were excluded (exclusion criteria are listed in the methods section). Seven studies were included in the review. Please cite this article in press as: Forker LJ, et al., Biomarkers of Tumour Radiosensitivity and Predicting Benefit from Radiotherapy, Clinical Oncology (2015), http://dx.doi.org/10.1016/j.clon.2015.06.002
L.J. Forker et al. / Clinical Oncology xxx (2015) 1e9
patients in the non-radiotherapy group (DMFS 80% radiosensitive versus 81% radioresistant, P ¼ 0.9425) [28]. 31 gene signature A 31 gene signature of radiosensitivity as defined by SF2 was derived from gene expression profiling data integrated from four different microarray platforms in NCI-60 panel cell lines [29]. This was then evaluated clinically in two independent cohorts of glioma patients. The 31 gene signature was used to classify patients as radiosensitive or radioresistant. In the Gene Expression Omnibus GSE16011 dataset of 276 glioma patients, radiosensitive patients had significantly longer overall survival than radioresistant patients (P < 0.0001). In the subgroup analysis, overall survival was superior for radiosensitive patients treated with radiotherapy and without radiotherapy in a univariate analysis, but only remained significant in the multivariate analysis in the radiotherapy treated group (P ¼ 0.0093 radiotherapy, P ¼ 0.202 no radiotherapy). In the second more homogeneous cohort of 463 glioblastoma multiforme patients from The Cancer Genome Atlas (TCGA), superior overall survival was observed in radiosensitive patients compared with radioresistant (P ¼ 0.000687) and remained significant in the multivariate analysis (P ¼ 0.033) [30]. The multivariate analysis for the GSE16011 cohort suggests that the gene signature may predict overall survival in radiotherapy treated glioma patients only. However, in the TCGA cohort, all patients were treated with radiotherapy. Therefore no firm conclusions can be drawn regarding the predictive versus prognostic value. MammaPrintÒ The ability of the 70 gene signature MammaPrintÒ to predict locoregional recurrence (LRR) has recently been evaluated in 1053 breast cancer patients treated with either breast-conserving therapy or mastectomy. Patients were classified by the signature as high or low risk. In the mastectomy group, the signature was evaluated separately in patients treated with and without post-mastectomy radiotherapy (PMRT). Those treated with PMRT and classified as high risk experienced more LRR at 10 years (LRR 11.0% high risk versus 3.1% low risk, P ¼ 0.004), whereas in patients who did not receive PMRT, classification as high or low risk did not affect LRR (LRR 12.9% high risk versus 9.2% low risk, P ¼ 0.302) [31]. Although the signature was predictive of LRR exclusively in radiotherapy treated patients in the mastectomy subgroup, high-risk classification by MammaPrintÒ was also shown to be an independent prognostic factor for LRR in the entire cohort and has already been proven to identify patients at high risk of distant metastasis [21]. Therefore, it is probably a purely prognostic marker. Danish Breast Cancer Cooperative Group radiotherapy profile Tramm et al. [32] identified seven genes associated with LRR in a training set of 191 patients with high-risk breast cancer treated with mastectomy and randomised between
5
PMRT and no PMRT in the DBCG82bc cohort. Patients were classified as high or low LRR risk according to the gene expression index. Radiotherapy significantly improved local control at 20 years in high-risk patients (LRR 57% no PMRT versus 12% PMRT, P < 0.0001), but was not beneficial in the low-risk patients (LRR 8% no PMRT versus 9% PMRT, P ¼ 0.93). The signature was then validated in 112 additional patients. Radiotherapy was beneficial in improving local control in high-risk patients (LRR 30% no PMRT versus 7% PMRT, P ¼ 0.003), but not low-risk patients (LRR 8% no PMRT versus 0% PMRT, P ¼ 0.30) [32]. Interferon-related DNA damage resistance signature Weichselbaum et al. [33] developed an interferon-related DNA damage resistance signature (IRDS) associated with resistance to chemotherapy or radiotherapy in cell lines. This was used to develop a seven gene signature in order to classify patients as IRDSþ (DNA damage resistant) or IRDSe (DNA damage sensitive). The IRDS was initially evaluated in 295 breast cancer patients and was shown to predict metastasis in patients treated with adjuvant chemotherapy versus nonchemotherapy treated patients. IRDSþ patients had higher locoregional failure rates than IRDSe patients after radiotherapy. However, there were too few events to evaluate the effect of IRDS in the non-radiotherapy treated group. In order to confirm its value as a therapy-specific marker, IRDS was then evaluated in three validation cohorts, one treated with adjuvant chemotherapy and/or radiotherapy (cohort A, n ¼ 292), one treated with adjuvant endocrine therapy but no other systemic therapy (cohort B, n ¼ 277) and one group who did not receive adjuvant systemic therapy (cohort C, n ¼ 286). In the first group, IRDSþ patients had shorter RFS than IRDSe patients; IRDS was also associated with metastasis-free survival (MFS) in those receiving adjuvant chemotherapy and with locoregional control in those treated with radiotherapy. There was no association with MFS for the other two cohorts and data regarding whether these groups received radiotherapy were not available. All the radiotherapy treated patients in cohort A received adjuvant chemotherapy. ‘Importance scores’ for the prediction of MFS and locoregional control were calculated for IRDS and other variables and compared in a meta-analysis including all patients. This showed a higher importance score for IRDS in predicting MFS in chemotherapy treated versus nonchemotherapy treated patients and a higher importance score of IRDS in predicting locoregional control in radiotherapy treated versus non-radiotherapy treated patients [33].
DNA Damage Response-related Biomarkers MRE11 MRE11 is a DDR protein that forms part of the MRN complex with RAD50 and NBS-1. MRN acts as a sensor of DSBs, recruits ATM and is involved in end resection during DSB repair [34].
Please cite this article in press as: Forker LJ, et al., Biomarkers of Tumour Radiosensitivity and Predicting Benefit from Radiotherapy, Clinical Oncology (2015), http://dx.doi.org/10.1016/j.clon.2015.06.002
6
L.J. Forker et al. / Clinical Oncology xxx (2015) 1e9
€derlund et al. [35] evaluated the expression of the DDR So proteins ATM, MRE11, RAD50 and NBS1 by immunohistochemistry (IHC) in tumour samples from 224 patients who underwent mastectomy for breast cancer and were randomised between adjuvant chemotherapy and adjuvant radiotherapy. In patients with moderate/strong MRN expression, local recurrence-free survival (LRFS) was higher for patients treated with radiotherapy than for those treated with chemotherapy (P ¼ 0.009), whereas in those with negative/weak MRN expression there was no difference in LRFS between chemotherapy and radiotherapy (P ¼ 0.73). Benefit in terms of local control in radiotherapy compared with chemotherapy was seen in patients with moderate/strong expression of the individual proteins RAD50, NBS1 and MRE11, but not for patients with negative/weak expression [35]. Choudhury et al. [36] assessed expression of the DDR proteins ATM, MRE11, RAD50, NBS1 and H2AX by IHC in two cohorts of muscle invasive bladder cancer (MIBC) patients treated with radical radiotherapy and a single cohort treated with radical cystectomy. In both radical radiotherapy groups, patients with high MRE11 showed improved cause-specific survival (CSS) at 3 years (cohort A high 68.7% versus low 43.1%, P ¼ 0.012; cohort B high 71.2% versus low 43.0%, P ¼ 0.02). However, in patients who underwent radical cystectomy, CSS was similar for high and low MRE11 expression (high 53.8% versus low 62.2%, P ¼ 0.46). There was no relationship with CSS for any of the other potential biomarkers assessed [36]. These results were replicated by Laurberg et al. [37] in a study that assessed expression of p16, ATM, TIP60, Rb, p53, ki67 and MRE11 by IHC in three cohorts of MIBC patients, one treated with radical radiotherapy with concurrent chemotherapy and two managed surgically. In 148 radical radiotherapy patients, those with high MRE11 expression had longer disease-specific survival (P ¼ 0.005). No value in the prediction of disease-specific survival was found for 273 cystectomy patients (P ¼ 0.96). TIP60 was predictive of outcome in cystectomy patients but not radiotherapy patients [37]. Most recently, Teo et al. [38] used next generation sequencing to explore the role of germline variants of MRE11A (the gene encoding MRE11) in MIBC. In 186 patients treated with radical radiotherapy, carriage of at least one of six rare variants was associated with worse 5 year CSS (P ¼ 0.009). The single nucleotide polymorphism (SNP) rs18005363 G > A was also associated with poor 5 year CSS (per allele hazard ratio 2.10, 95% confidence interval 1.34e3.28, P ¼ 0.001); 5 year CSS by genotype was 58.3% GG, 42.0% GA and 0% AA. No association of this SNP with CSS was seen for 256 cystectomy patients (5 year CSS 59.7% GG, 54.0% GA and 75.0% AA, P ¼ 0.89) [38]. AIMP3 AIMP3 is a tumour suppressor gene that acts as an upstream regulator of p53 in response to DNA damage. Gurung et al. [39] showed low AIMP3 expression in MIBC and evaluated AIMP3 and p53 for their ability to predict radiotherapy outcomes. AIMP3 expression was defined as
positive or negative by IHC. In 217 MIBC patients treated with radical radiotherapy carbogen and nicotinamide enrolled in the BCON study, the median survival was superior in AIMP3-positive patients (67.9 months positive versus 21.5 months negative, P ¼ 0.002). In 151 MIBC patients treated with cystectomy, AIMP3 expression did not correlate with survival (P ¼ 0.70). No interaction between AIMP3 expression and either BCON trial arm was found and p53 had no association with survival in radiotherapy or cystectomy cohorts [39]. NBN NBN is the gene encoding NBS-1, which also forms part of the MRN complex [34]. Berlin et al. [40] determined copy number variation in the DDR genes MRE11A, RAD50, NBN, ATM, ATR and PRKDC in pre-treatment prostate cancer biopsies. Variation in NBN copy number was the most common abnormality observed and was associated with genomic instability [40]. In 139 localised prostate cancer patients treated with radical image-guided radiotherapy, NBN gain was a significant predictor of biochemical relapse-free rate at 5 years (46% NBN gain versus 77% NBN neutral, P ¼ 0.00067) and remained significant in a multivariate analysis. In a control cohort of 154 localised prostate cancer patients treated with radical prostatectomy, NBN gain was not associated with biochemical relapse-free rate (61% NBN gain versus 77% NBN neutral, P ¼ 0.15). No association was found for any of the other genes assessed [40]. BRE BRE is a member of the BRCA1 complex, which is important in DSB repair. Noordermeer et al. [41] used quantitative PCR (QPCR) to determine BRE expression in 229 breast cancer patients. For the whole cohort, BRE expression was not prognostic. In 169 patients treated with adjuvant radiotherapy, high BRE expression correlated with improved disease-free survival (P ¼ 0.004) and remained significant on multivariate analysis. High BRE expression initially seemed to predict poor disease-free survival in non-radiotherapy treated patients, but did not remain significant on multivariate analysis. High BRE expression was a favourable prognostic factor in a larger cohort of 2324 breast cancer patients, but could not be assessed for treatment-specific predictive value [41].
Discussion Limitations of Radiotherapy Biomarkers Identified The literature search highlighted that despite many citations identified, few studies have been undertaken in which a radiotherapy biomarker had undergone robust clinical assessment in over 100 patients and was compared between radiotherapy and non-radiotherapy treated groups to confirm its value as a predictive rather than prognostic marker. Ten biomarkers were reported as being predictive of radiotherapy benefit, five were gene signatures and five were DDR-related markers.
Please cite this article in press as: Forker LJ, et al., Biomarkers of Tumour Radiosensitivity and Predicting Benefit from Radiotherapy, Clinical Oncology (2015), http://dx.doi.org/10.1016/j.clon.2015.06.002
L.J. Forker et al. / Clinical Oncology xxx (2015) 1e9
RSI is the most extensively clinically investigated radiosensitivity signature, having been assessed in four different tumour types in five independent cohorts [27,28]. However, the rectal, oesophageal and head and neck cancer datasets represented preliminary work and so included low numbers of patients, findings were not validated in a second cohort or compared with a group of non-radiotherapy treated controls. RSI predicted improved RFS exclusively in patients who received adjuvant radiotherapy in two larger independent cohorts of over 500 breast cancer patients in total. The 31 gene signature investigated in glioma was found to be predictive of longer overall survival in radiotherapy treated patients only in the GSE16011 cohort and in the TCGA cohort in which all patients received radiotherapy [30]. Further work is necessary to conclude that this is definitely a predictive rather than a prognostic marker, as it has only been evaluated in one cohort that allowed comparison between radiotherapy and non-radiotherapy treated patients and was associated with overall survival in both groups on univariate analysis. Neither RSI nor the 31 gene signature has been investigated in a randomised controlled trial (RCT). The other three signatures were investigated for value in predicting benefit from adjuvant radiotherapy in breast cancer. Weichselbaum et al. [33] concluded that the IRDS signature predicted benefit from radiotherapy. However, there is stronger evidence for its value in predicting the response to adjuvant chemotherapy. All of the radiotherapy treated patients in the validation cohort also received chemotherapy and so the poorer locoregional control seen in the IRDSþ group may have been due to chemoresistance rather than radioresistance. It is also unclear as to whether this has truly been assessed in radiotherapy versus nonradiotherapy treated patients, as information was not available regarding radiotherapy for two of the cohorts included in the meta-analysis. MammaPrintÒ is probably a prognostic rather than a predictive biomarker and the Danish Breast Cancer Cooperative Group radiotherapy profile requires further validation in a larger cohort as part of an RCT. The most promising DDR-related biomarker of radiotherapy outcomes identified was MRE11. High tumour MRE11 expression has been shown to correlate with improved LRFS in breast cancer only in patients who received adjuvant radiotherapy [35] and with CSS in MIBC in radical radiotherapy but not cystectomy patients. In MIBC these findings have been validated in a large number of patients; two independent radiotherapy cohorts were included in the first study published [36] and findings have been replicated in another cohort by a different group [37]. It is reasonable to conclude that MRE11 may be a predictive biomarker of benefit from radiotherapy and work is ongoing to assess this in patients previously recruited as part of the phase III RCTs BCON [42] and BC2001 [43]. A germline SNP in MRE11A has also been shown to predict outcome in radiotherapy patients but not cystectomy patients in MIBC, but is yet to be validated in an independent cohort [38]. This highlights a major problem, namely that germline blood
7
samples have not been routinely collected in radiotherapybased studies to date. Other DDR-related biomarkers identified as predictive of outcome exclusively in radiotherapy treated patients were AIMP3 in MIBC [39], NBN in prostate cancer [40] and BRE in breast cancer [41], but these findings have not yet been replicated. Although some of these biomarkers show promise as being predictive of radiotherapy outcomes, particularly RSI and MRE11, none has been assessed in a prospective RCT, which would be the gold standard approach to validate a predictive marker. Therefore, further studies are needed before they can be introduced into standard practice. Challenges in Implementing Biomarker-driven Radiotherapy Despite the growing field of cancer biomarker research and numerous papers published in the literature, no predictive biomarkers of radiotherapy outcome have transitioned to use as the standard of care. One difficulty is in providing levels of evidence high enough to justify routine clinical use. This is particularly difficult for predictive markers, which must be assessed in relation to outcomes in a therapy-specific manner [11,44]. Low patient numbers and lack of a control population were common reasons for exclusion of biomarkers from this review. Predictive biomarker validation requires an RCT in order to ensure the patient groups are comparable and so the collection of samples for translational research is extremely important. Ideally a prospective RCT in which biomarker validation is included in the trial design should be used, but retrospective analysis of previous RCTs is also acceptable [11]. None of the biomarkers identified by this review was compared between patients who received or did not receive radiotherapy in an RCT. Assessment of a radiotherapy biomarker in this way is particularly difficult as RCTs of radical radiotherapy versus non-radiotherapy radical treatments are rare. For example, in MIBC, radical radiotherapy and surgery are said to be comparable in efficacy on the basis of retrospective data [45], but an attempt to evaluate this in an RCT was aborted [46]. If a predictive biomarker were to be integrated into routine practice it would require a convenient, reproducible assay capable of delivering quick results. The REMARK guidelines for reporting of biomarker studies specify the importance of inclusion of a detailed assay method [44,47]. IHC is often used in the assessment of biomarkers in tumour samples, but can give variable results due to inconsistencies in methodology and interpretation [48,49]. The evaluation of biomarkers at the level of RNA may be more easily reproducible. However, for some markers, posttranscriptional modifications mean that mRNA levels and protein levels do not always correlate [50]. SNP genotyping for germline mutations, as used in the study of MRE11A, would be attractive as this could be carried out on a peripheral blood sample. The studies of MRE11 expression used IHC, which would be challenging to implement in routine practice, as this would require rigorous guidelines for standardisation across laboratories, although this has
Please cite this article in press as: Forker LJ, et al., Biomarkers of Tumour Radiosensitivity and Predicting Benefit from Radiotherapy, Clinical Oncology (2015), http://dx.doi.org/10.1016/j.clon.2015.06.002
8
L.J. Forker et al. / Clinical Oncology xxx (2015) 1e9
been achievable for other tests, such as HER-2 in breast cancer [51].
Conclusion Currently there are no biomarkers predictive of radiosensitivity or radiotherapy benefit used as standard of care. There are several on the horizon that show promise but require further clinical assessment. Predictive biomarkers for targeted agents have revolutionised practice in some areas of systemic treatment but these only affect a small proportion of cancer patients. As radiotherapy is such a commonly utilised treatment modality the radiosensitivity biomarkers identified have the potential to improve treatment outcomes for very large numbers of patients and therefore would have a huge effect on oncology practice.
References [1] Ferlay J, Soerjomataram I, Ervik M, et al. GLOBOCAN 2012 v1.0, Cancer Incidence and Mortality Worldwide: IARC CancerBase No. 11 [database on the Internet]. Lyon: International Agency for Research on Cancer. Available at: http://globocan.iarc.fr; 2013. [2] Horwich A, Parker C, de Reijke T, Kataja V. Prostate cancer: ESMO clinical practice guidelines for diagnosis, treatment and follow-up. Ann Oncol 2013;24(Suppl. 6):vi106evi114. [3] Vansteenkiste J, De Ruysscher D, Eberhardt WEE, et al. Early and locally advanced non-small-cell lung cancer (NSCLC): ESMO clinical practice guidelines for diagnosis, treatment and follow-up. Ann Oncol 2013;24(Suppl. 6):vi89evi98. [4] Senkus E, Kyriakides S, Penault-Llorca F, et al. Primary breast cancer: ESMO clinical practice guidelines for diagnosis, treatment and follow-up. Ann Oncol 2013;24(Suppl. 6):vi7evi23. [5] Glimelius B, Tiret E, Cervantes A, Arnold D. Rectal cancer: ESMO clinical practice guidelines for diagnosis, treatment and follow-up. Ann Oncol 2013;24(Suppl. 6):vi81evi88. [6] Lutz ST, Jones J, Chow E. Role of radiation therapy in palliative care of the patient with cancer. J Clin Oncol 2014;32(26): 2913e2919. [7] Delaney G, Jacob S, Featherstone C, Barton M. The role of radiotherapy in cancer treatment. Cancer 2005;104(6): 1129e1137. [8] Staffurth J. A review of the clinical evidence for intensitymodulated radiotherapy. Clin Oncol 2010;22(8):643e657. [9] Hanahan D, Weinberg RA. Hallmarks of cancer: the next generation. Cell 2011;144(5):646e674. [10] Burrell RA, McGranahan N, Bartek J, Swanton C. The causes and consequences of genetic heterogeneity in cancer evolution. Nature 2013;501(7467):338e345. [11] Mandrekar SJ, Sargent DJ. Clinical trial designs for predictive biomarker validation: theoretical considerations and practical challenges. J Clin Oncol 2009;27(24):4027e4034. [12] Polley M-YC, Freidlin B, Korn EL, Conley BA, Abrams JS, McShane LM. Statistical and practical considerations for clinical evaluation of predictive biomarkers. J Natl Cancer Inst 2013;105(22):1677e1683. [13] Slamon DJ, Leyland-Jones B, Shak S, et al. Use of chemotherapy plus a monoclonal antibody against HER2 for metastatic breast cancer that overexpresses HER2. N Engl J Med 2001;344(11):783e792.
s J, Kim S-B, et al. Pertuzumab plus trastuzu[14] Baselga J, Corte mab plus docetaxel for metastatic breast cancer. N Engl J Med 2012;366(2):109e119. [15] Verma S, Miles D, Gianni L, et al. Trastuzumab emtansine for HER2-positive advanced breast cancer. N Engl J Med 2012;367(19):1783e1791. [16] Takano T, Ohe Y, Sakamoto H, et al. Epidermal growth factor receptor gene mutations and increased copy numbers predict gefitinib sensitivity in patients with recurrent nonesmall-cell lung cancer. J Clin Oncol 2005;23(28):6829e6837. [17] Mitsudomi T, Morita S, Yatabe Y, et al. Gefitinib versus cisplatin plus docetaxel in patients with non-small-cell lung cancer harbouring mutations of the epidermal growth factor receptor (WJTOG3405): an open label, randomised phase 3 trial. Lancet Oncol 2010;11(2):121e128. [18] Shaw AT, Yeap BY, Solomon BJ, et al. Effect of crizotinib on overall survival in patients with advanced non-small-cell lung cancer harbouring ALK gene rearrangement: a retrospective analysis. Lancet Oncol 2011;12(11):1004e1012. [19] Solomon BJ, Mok T, Kim D-W, et al. First-line crizotinib versus chemotherapy in ALK-positive lung cancer. N Engl J Med 2014;371(23):2167e2177. [20] West CM, Davidson SE, Roberts SA, Hunter RD. The independence of intrinsic radiosensitivity as a prognostic factor for patient response to radiotherapy of carcinoma of the cervix. Br J Cancer 1997;76(9):1184e1190. [21] van de Vijver MJ, He YD, van’t Veer LJ, et al. A gene-expression signature as a predictor of survival in breast cancer. N Engl J Med 2002;347(25):1999e2009. [22] Paik S, Shak S, Tang G, et al. A multigene assay to predict recurrence of tamoxifen-treated, node-negative breast cancer. N Engl J Med 2004;351(27):817e826. [23] Eustace A, Mani N, Span PN, et al. A 26-gene hypoxia signature predicts benefit from hypoxia-modifying therapy in laryngeal cancer but not bladder cancer. Clin Cancer Res 2013;19(17):4879e4888. [24] Wouters BG, Begg AC. Irradiation-induced damage and the DNA damage response. In: Joiner M, van der Kogel A, editors. Basic clinical radiobiology. Great Britain: CRC Press; 2009. p. 11e26. € hne M, Rief N, et al. A pathway of double-strand [25] Riballo E, Ku break rejoining dependent upon ATM, Artemis and proteins locating to g-H2AX foci. Mol Cell 2004;16(5):715e724. [26] Torres-Roca JF, Eschrich S, Zhao H, et al. Prediction of radiation sensitivity using a gene expression classifier. Cancer Res 2005;65(16):7169e7176. [27] Eschrich SA, Pramana J, Zhang H, et al. A gene expression model of intrinsic tumor radiosensitivity: prediction of response and prognosis after chemoradiation. Int J Radiat Oncol Biol Phys 2009;75(2):489e496. [28] Eschrich SA, Fulp WJ, Pawitan Y, et al. Validation of a radiosensitivity molecular signature in breast cancer. Clin Cancer Res 2012;18(18):5134e5143. [29] Kim H, Kim S, Kim S, et al. Identification of a radiosensitivity signature using integrative metaanalysis of published microarray data for NCI-60 cancer cells. BMC Genomics 2012;13(1):348. [30] Meng J, Li P, Zhang Q, Yang Z, Fu S. A radiosensitivity gene signature in predicting glioma prognostic via EMT pathway. Oncotarget 2014;5(13):4683e4693. [31] Drukker CA, Elias SG, Nijenhuis MV, et al. Gene expression profiling to predict the risk of locoregional recurrence in breast cancer: a pooled analysis. Breast Cancer Res Treat 2014;148(3):599e613. [32] Tramm T, Mohammed H, Myhre S, et al. Development and validation of a gene profile predicting benefit of
Please cite this article in press as: Forker LJ, et al., Biomarkers of Tumour Radiosensitivity and Predicting Benefit from Radiotherapy, Clinical Oncology (2015), http://dx.doi.org/10.1016/j.clon.2015.06.002
L.J. Forker et al. / Clinical Oncology xxx (2015) 1e9
[33]
[34] [35]
[36]
[37]
[38]
[39]
[40]
[41]
postmastectomy radiotherapy in patients with high-risk breast cancer: a study of gene expression in the DBCG82bc cohort. Clin Cancer Res 2014;20(20):5272e5280. Weichselbaum RR, Ishwaran H, Yoon T, et al. An interferonrelated gene signature for DNA damage resistance is a predictive marker for chemotherapy and radiation for breast cancer. Proc Natl Acad Sci USA 2008;105(47):18490e18495. Harper JW, Elledge SJ. The DNA damage response: ten years after. Mol Cell 2007;28(5):739e745. € derlund K, St € ld B, So al O, Skoog L, Rutqvist LE, Nordenskjo Askmalm MS. Intact Mre11/Rad50/Nbs1 complex predicts good response to radiotherapy in early breast cancer. Int J Radiat Oncol Biol Phys 2007;68(1):50e58. Choudhury A, Nelson LD, Teo MTW, et al. MRE11 expression is predictive of cause-specific survival following radical radiotherapy for muscle invasive bladder cancer. Cancer Res 2010;70(18):7017e7026. Laurberg JR, Brems-Eskildsen AS, Nordentoft I, et al. Expression of TIP60 (tat-interactive protein) and MRE11 (meiotic recombination 11 homolog) predict treatment-specific outcome of localised invasive bladder cancer. BJU Int 2012;110(11c):E1228eE1236. Teo MTW, Dyrskjøt L, Nsengimana J, et al. Next-generation sequencing identifies germline MRE11A variants as markers of radiotherapy outcomes in muscle-invasive bladder cancer. Ann Oncol 2014;25(4):877e883. Gurung PMS, Veerakumarasivam A, Williamson M, et al. Loss of expression of the tumour suppressor gene AIMP3 predicts survival following radiotherapy in muscle-invasive bladder cancer. Int J Cancer 2015;136(3):709e720. Berlin A, Lalonde E, Sykes J, et al. NBN gain is predictive for adverse outcome following image-guided radiotherapy for localized prostate cancer. Oncotarget 2014;5(22): 11081e11090. Noordermeer SM, Wennemers M, Bergevoet SM, et al. Expression of the BRCA1 complex member BRE predicts
[42]
[43]
[44]
[45]
[46]
[47]
[48]
[49]
[50]
[51]
9
disease free survival in breast cancer. Breast Cancer Res Treat 2012;135(1):125e133. Hoskin PJ, Rojas AM, Bentzen SM, Saunders MI. Radiotherapy with concurrent carbogen and nicotinamide in bladder carcinoma. J Clin Oncol 2010;28(33):4912e4918. James ND, Hussain SA, Hall E, et al. Radiotherapy with or without chemotherapy in muscle-invasive bladder cancer. N Engl J Med 2012;366(16):1477e1488. Altman DG, McShane LM, Sauerbrei W, Taube SE. Reporting recommendations for tumor marker prognostic studies (REMARK): explanation and elaboration. PLoS Med 2012;9(5):e1001216. Kotwal S, Choudhury A, Johnston C, Paul AB, Whelan P, Kiltie AE. Similar treatment outcomes for radical cystectomy and radical radiotherapy in invasive bladder cancer treated at a United Kingdom specialist treatment center. Int J Radiat Oncol Biol Phys 2008;70(2):456e463. Huddart RA, Hall E, Lewis R, Birtle A, on behalf of the STMG. Life and death of SPARE (Selective Bladder Preservation Against Radical Excision): reflections on why the spare trial closed. BJU Int 2010;106(6):753e755. McShane LM, Altman DG, Sauerbrei W, et al. Reporting recommendations for tumor marker prognostic studies (REMARK). Natl Clin Prac Oncol 2005;2(8):416e422. McShane LM, Aamodt R, Cordon-Cardo C, et al. Reproducibility of p53 immunohistochemistry in bladder tumors. Clin Cancer Res 2000;6(5):1854e1864. Bussolati G, Leonardo E. Technical pitfalls potentially affecting diagnoses in immunohistochemistry. J Clin Pathol 2008; 61(11):1184e1192. Martin RM, Kerr M, Teo MTW, et al. Post-transcriptional regulation of MRE11 expression in muscle-invasive bladder tumours. Oncotarget 2014;5(4):993e1003. Rakha EA, Pinder SE, Bartlett JMS, et al. Updated UK recommendations for HER2 assessment in breast cancer. J Clin Pathol 2014;68(2):93e99.
Please cite this article in press as: Forker LJ, et al., Biomarkers of Tumour Radiosensitivity and Predicting Benefit from Radiotherapy, Clinical Oncology (2015), http://dx.doi.org/10.1016/j.clon.2015.06.002
Clinical Oncology journal homepage: www.clinicaloncologyonline.net
Overview
Biomarkers of Tumour Radiosensitivity and Predicting Benefit from Radiotherapy L.J. Forker *y, A. Choudhury *y, A.E. Kiltie z * Institute
of Cancer Sciences, The University of Manchester, Manchester Academic Health Science Centre, The Christie NHS Foundation Trust, Manchester, UK y Department of Clinical Oncology, The Christie NHS Foundation Trust, Manchester, UK z CRUK/MRC Oxford Institute for Radiation Oncology, Department of Oncology, University of Oxford, Old Road Campus Research Building, Headington, Oxford, UK Received 5 February 2015; accepted 2 June 2015
Abstract Radiotherapy is an essential component of treatment for more than half of newly diagnosed cancer patients. The response to radiotherapy varies widely between individuals and although advances in technology have allowed the adaptation of radiotherapy fields to tumour anatomy, it is still not possible to tailor radiotherapy based on tumour biology. A biomarker of intrinsic radiosensitivity would be extremely valuable for individual dosing, aiding decision making between radical treatment options and avoiding toxicity of neoadjuvant or adjuvant radiotherapy in those unlikely to benefit. This systematic review summarises the current evidence for biomarkers under investigation as predictors of radiotherapy benefit. Only 10 biomarkers were identified as having been evaluated for their radiotherapy-specific predictive value in over 100 patients in a clinical setting, highlighting that despite a rich literature there were few highquality studies for inclusion. The most extensively studied radiotherapy predictive biomarkers were the radiosensitivity index and MRE11; however, neither has been evaluated in a randomised controlled trial. Although these biomarkers show promise, there is not enough evidence to justify their use in routine practice. Further validation is needed before biomarkers can fulfil their potential and predict treatment outcomes for large numbers of patients. Ó 2015 Published by Elsevier Ltd on behalf of The Royal College of Radiologists.
Key words: DNA damage response; molecular signature; predictive biomarker; radiosensitivity; radiotherapy
Statement of Search Strategies Used and Sources of Information A literature search of PubMed for molecular signatures predictive of radiotherapy benefit was conducted using the terms (predict OR prediction OR predictive OR predictor OR predicts) AND (radiotherapy OR radiosensitivity OR chemoradiotherapy) AND (‘radiosensitivity index’ OR gene signature OR molecular signature OR gene expression profile) and yielded 122 results. A second literature search of PubMed for biomarkers related to DNA damage response predictive of radiotherapy benefit using the terms (predict
Author for correspondence: A. Choudhury, Institute of Cancer Sciences, The University of Manchester, Manchester Academic Health Science Centre, The Christie NHS Foundation Trust, Wilmslow Road, Manchester M20 4BX, UK. Tel: þ44-161-446-3000; Fax: þ44-161-446-3977. E-mail address: [email protected] (A. Choudhury).
OR prediction OR predictive OR predictor OR predicts) AND (radiotherapy OR radiosensitivity OR chemoradiotherapy) AND (biomarker OR molecular OR signature) AND (DNA repair OR DNA damage OR DNA OR ‘DNA damage response’) found 476 results. Searches were supplemented by hand searching of reference lists of relevant studies and reviews. Abstracts were reviewed for all relevant titles and full papers obtained if necessary. Citations were excluded if they were not original research articles or not relevant to cancer biomarker research. Biomarkers predicting benefit from chemotherapy, hypoxia modifying therapy, targeted agents and endocrine therapy were excluded, as were those concerning diagnostic or prognostic markers and those predictive of normal tissue toxicity. Citations relating to imaging markers were excluded. Pre-clinical work, studies with small patient numbers (<100) and studies with no control population were also excluded. Only reports published in English up to January 2015 week 1 were included.
http://dx.doi.org/10.1016/j.clon.2015.06.002 0936-6555/Ó 2015 Published by Elsevier Ltd on behalf of The Royal College of Radiologists.
Please cite this article in press as: Forker LJ, et al., Biomarkers of Tumour Radiosensitivity and Predicting Benefit from Radiotherapy, Clinical Oncology (2015), http://dx.doi.org/10.1016/j.clon.2015.06.002
2
L.J. Forker et al. / Clinical Oncology xxx (2015) 1e9
Introduction Radiotherapy in Cancer Treatment In 2012 there were 14.1 million new cancer diagnoses worldwide, the most common being lung (13%), breast (11.9%), colorectal (9.7%) and prostate (7.9%) [1]. Radiotherapy is recommended as a key component of curative management of these cancers, as either neoadjuvant, definitive or adjuvant treatment [2e5] and is essential in palliative oncology care [6]. Overall more than half of newly diagnosed cancer patients will require radiotherapy [7]. Although recent technological advances in radiotherapy planning and delivery techniques have led to more individualised radiotherapy based on the anatomy of the tumour with subsequent improved therapeutic ratio [8], the adaptation of radiotherapy with respect to tumour biology has significant potential to contribute to further therapeutic gain. Predictive Cancer Biomarkers Cancer is characterised by considerable genetic and epigenetic heterogeneity driven by genomic instability [9,10]. It is therefore not surprising that even patients diagnosed with the same cancer type vary widely in the natural progression of their disease and response to treatment. An era of personalised cancer medicine in which biomarkers can be used to tailor treatment to each specific patient is a major goal in oncology. Prognostic biomarkers provide information regarding disease outcome regardless of the treatment received and predictive biomarkers determine which patients will probably derive benefit from a specific therapy [11,12]. Examples of the few predictive biomarkers that have successfully transitioned to routine clinical use are those predicting response to targeted drugs, such as HER-2 in selecting breast cancer patients for HER-2 targeted therapies [13e15] and EGFR and ALK mutations in predicting response to tyrosine kinase inhibitors in non-small cell lung cancer [16e19]. These biomarkers have had a dramatic effect on current practice for selected populations of cancer patients. Radiosensitivity Biomarkers Potential applications of a radiosensitivity biomarker include tailoring of the dose depending on tumour biology, aiding decision making between radical radiotherapy and radical surgery where both are viable options and avoiding delay of definitive surgery for those unlikely to benefit from neoadjuvant radiotherapy or chemoradiotherapy. Radiosensitivity biomarkers may also be useful in predicting normal tissue toxicity, but this will be discussed in another review. Many factors are known to influence tumour response to irradiation, including total dose, fractionation, tumour potential doubling time, hypoxia and intrinsic radiosensitivity.
Fractionation and hypoxia will be discussed in separate reviews. Traditional laboratory tests to determine intrinsic radiosensitivity, such as measuring ex vivo surviving fraction at 2 Gy (SF2) by clonogenic survival assay, have been correlated with clinical outcome [20], but are not practical for routine use. Therefore alternative strategies must be sought. The development of high throughput molecular profiling techniques has led to the identification of gene expression signatures in tumours. Signatures such as Oncotype DxÒ and MammaPrintÒ have been successful in stratification for recurrence risk and aiding decision making regarding adjuvant chemotherapy in breast cancer [21,22] and a hypoxia signature has been shown to predict benefit from concurrent hypoxia modification with radiotherapy in laryngeal cancer [23]. This may also be an attractive approach for the development of a biomarker of intrinsic radiosensitivity. Additionally molecules involved in DNA damage response (DDR) signalling pathways (Figure 1), which sense and repair DNA damage [24], are excellent candidates for evaluation as radiosensitivity biomarkers, as cells with a defective DDR have less ability to repair lethal radiation-induced DNA double-strand breaks (DSBs) and are therefore more sensitive to irradiation, as seen in radiosensitivity syndromes such as ataxia telangiectasia [25]. Aims of Review Cancer biomarker research is an ever-expanding field with large numbers of publications, including numerous pre-clinical studies. In order to focus on the most promising radiosensitivity biomarkers, this review will concentrate on molecular signatures and DDR-related biomarkers that have been evaluated for their ability to predict benefit from radiotherapy in a clinical setting.
Materials and Methods Literature Searches A literature search of PubMed for molecular signatures predictive of radiotherapy benefit was conducted using the terms (predict OR prediction OR predictive OR predictor OR predicts) AND (radiotherapy OR radiosensitivity OR chemoradiotherapy) AND (‘radiosensitivity index’ OR gene signature OR molecular signature OR gene expression profile) and yielded 122 results. A second literature search of PubMed for biomarkers related to DDR predictive of radiotherapy benefit using the terms (predict OR prediction OR predictive OR predictor OR predicts) AND (radiotherapy OR radiosensitivity OR chemoradiotherapy) AND (biomarker OR molecular OR signature) AND (DNA repair OR DNA damage OR DNA OR ‘DNA damage response’) found 476 results. Searches were supplemented by hand searching of reference lists of relevant studies and reviews. Abstracts were reviewed for all relevant titles and full papers obtained if necessary.
Please cite this article in press as: Forker LJ, et al., Biomarkers of Tumour Radiosensitivity and Predicting Benefit from Radiotherapy, Clinical Oncology (2015), http://dx.doi.org/10.1016/j.clon.2015.06.002
L.J. Forker et al. / Clinical Oncology xxx (2015) 1e9
3
Fig 1. Response to DNA double-strand breaks (DSBs) induced by ionising radiation. The MRN complex (MRE11, RAD50 and NBS1) and Ku70/ Ku80 act as sensors of DSBs and activate ATM and DNA-PKcs, respectively. These phosphorylate the histone H2AX, amplifying the signal and recruiting molecules involved in the DNA damage response (DDR) to form irradiation-induced foci (IRIF). ATM phosphorylates p53 and MDM2 (a protein that normally binds p53 tagging it for destruction) causing dissociation of p53 from MDM2. p53 acts as a transcription factor, which upregulates pro-apoptotic genes and p21 (a cyclin-dependent kinase inhibitor), leading to apoptosis or G1 phase arrest, respectively. ATM also phosphorylates CHK2 to induce s phase arrest. Homologous recombination and non-homologous end joining are two mechanisms of DSB repair. In homologous recombination, MRE11 is involved in end resection to create single-stranded regions before strand invasion of homologous DNA mediated by RAD51 and XRCC2/3, which is used as a template for repair. BRCA 1 and 2 have multiple roles in homologous recombination. In nonhomologous end joining, Ku70 and Ku80 bind broken DNA ends and recruit DNA-PKcs, which phosphorylates other involved proteins. ARTEMIS processes the ends prior to ligation by LIG4, which is aided by XRCC4 [24].
Exclusion Criteria Citations were excluded if they were not original research articles or not relevant to cancer biomarker research. Biomarkers predicting benefit from chemotherapy, hypoxia modifying therapy, targeted agents and endocrine therapy were excluded, as were those concerning diagnostic or prognostic markers and those predictive of normal tissue toxicity. Citations relating to imaging markers were excluded. Pre-clinical work, studies with small patient numbers (<100) and studies with no control population were also excluded. Only reports published in English up to January 2015 week 1 were included. Figures 2 and 3 show flow diagrams of citations identified and reasons for exclusion.
Results Molecular Signatures Radiosensitivity index Torres-Roca et al. [26] were the first to show that a gene expression profile could be used to predict SF2 in cancer cell lines of numerous types [26]. A 10 gene radiosensitivity signature was developed in 48 cancer cell lines from the National Cancer Institute 60 (NCI-60) panel. Radiosensitivity was modelled as cell survival post-irradiation as a function of expression of these genes to give the
radiosensitivity index (RSI), where a lower RSI indicates higher radiosensitivity [27]. RSI was initially evaluated clinically in 14 rectal cancer patients treated with neoadjuvant chemoradiotherapy and 12 oesophageal cancer patients treated with preoperative concurrent chemoradiotherapy followed by planned esophagectomy. RSI was lower for patients who responded (defined as decrease in T stage or clinical complete response). In 92 head and neck cancer patients treated with radiotherapy and concurrent cisplatin-based chemotherapy patients with lower RSI had superior locoregional control at 2 years (locoregional control 86% radiosensitive versus 61% radioresistant, P ¼ 0.05) [27]. RSI has also been evaluated retrospectively in breast cancer in two independent cohorts. In the first cohort (n ¼ 159), radiotherapy treated patients classified as radiosensitive by RSI had improved recurrence-free survival (RFS) at 5 years (RFS 95% radiosensitive versus 75% radioresistant, P ¼ 0.0212), but there was no difference in RFS between radiosensitive and radioresistant patients in those not treated with radiotherapy (RFS 71% radiosensitive versus 77% radioresistant, P ¼ 0.6744). Most relapses were distant metastases. In a second more homogeneous cohort of 344 patients who were lymph node negative and received no adjuvant systemic treatment, radiosensitive patients had superior distant metastasis-free survival (DMFS) in the radiotherapy treated group (DMFS 77% radiosensitive versus 64% radioresistant, P ¼ 0.0409), but DMFS was similar for radiosensitive and radioresistant
Please cite this article in press as: Forker LJ, et al., Biomarkers of Tumour Radiosensitivity and Predicting Benefit from Radiotherapy, Clinical Oncology (2015), http://dx.doi.org/10.1016/j.clon.2015.06.002
4
L.J. Forker et al. / Clinical Oncology xxx (2015) 1e9
Fig 2. Flow diagram for the literature search for molecular signature biomarkers predictive of radiotherapy benefit. In total, 122 citations were identified and 117 of these were excluded (exclusion criteria are listed in the methods section). Five studies were included in the review.
Fig 3. Flow diagram for the literature search for DNA damage response-related biomarkers predictive of radiotherapy benefit. In total, 476 citations were identified and 469 of these were excluded (exclusion criteria are listed in the methods section). Seven studies were included in the review. Please cite this article in press as: Forker LJ, et al., Biomarkers of Tumour Radiosensitivity and Predicting Benefit from Radiotherapy, Clinical Oncology (2015), http://dx.doi.org/10.1016/j.clon.2015.06.002
L.J. Forker et al. / Clinical Oncology xxx (2015) 1e9
patients in the non-radiotherapy group (DMFS 80% radiosensitive versus 81% radioresistant, P ¼ 0.9425) [28]. 31 gene signature A 31 gene signature of radiosensitivity as defined by SF2 was derived from gene expression profiling data integrated from four different microarray platforms in NCI-60 panel cell lines [29]. This was then evaluated clinically in two independent cohorts of glioma patients. The 31 gene signature was used to classify patients as radiosensitive or radioresistant. In the Gene Expression Omnibus GSE16011 dataset of 276 glioma patients, radiosensitive patients had significantly longer overall survival than radioresistant patients (P < 0.0001). In the subgroup analysis, overall survival was superior for radiosensitive patients treated with radiotherapy and without radiotherapy in a univariate analysis, but only remained significant in the multivariate analysis in the radiotherapy treated group (P ¼ 0.0093 radiotherapy, P ¼ 0.202 no radiotherapy). In the second more homogeneous cohort of 463 glioblastoma multiforme patients from The Cancer Genome Atlas (TCGA), superior overall survival was observed in radiosensitive patients compared with radioresistant (P ¼ 0.000687) and remained significant in the multivariate analysis (P ¼ 0.033) [30]. The multivariate analysis for the GSE16011 cohort suggests that the gene signature may predict overall survival in radiotherapy treated glioma patients only. However, in the TCGA cohort, all patients were treated with radiotherapy. Therefore no firm conclusions can be drawn regarding the predictive versus prognostic value. MammaPrintÒ The ability of the 70 gene signature MammaPrintÒ to predict locoregional recurrence (LRR) has recently been evaluated in 1053 breast cancer patients treated with either breast-conserving therapy or mastectomy. Patients were classified by the signature as high or low risk. In the mastectomy group, the signature was evaluated separately in patients treated with and without post-mastectomy radiotherapy (PMRT). Those treated with PMRT and classified as high risk experienced more LRR at 10 years (LRR 11.0% high risk versus 3.1% low risk, P ¼ 0.004), whereas in patients who did not receive PMRT, classification as high or low risk did not affect LRR (LRR 12.9% high risk versus 9.2% low risk, P ¼ 0.302) [31]. Although the signature was predictive of LRR exclusively in radiotherapy treated patients in the mastectomy subgroup, high-risk classification by MammaPrintÒ was also shown to be an independent prognostic factor for LRR in the entire cohort and has already been proven to identify patients at high risk of distant metastasis [21]. Therefore, it is probably a purely prognostic marker. Danish Breast Cancer Cooperative Group radiotherapy profile Tramm et al. [32] identified seven genes associated with LRR in a training set of 191 patients with high-risk breast cancer treated with mastectomy and randomised between
5
PMRT and no PMRT in the DBCG82bc cohort. Patients were classified as high or low LRR risk according to the gene expression index. Radiotherapy significantly improved local control at 20 years in high-risk patients (LRR 57% no PMRT versus 12% PMRT, P < 0.0001), but was not beneficial in the low-risk patients (LRR 8% no PMRT versus 9% PMRT, P ¼ 0.93). The signature was then validated in 112 additional patients. Radiotherapy was beneficial in improving local control in high-risk patients (LRR 30% no PMRT versus 7% PMRT, P ¼ 0.003), but not low-risk patients (LRR 8% no PMRT versus 0% PMRT, P ¼ 0.30) [32]. Interferon-related DNA damage resistance signature Weichselbaum et al. [33] developed an interferon-related DNA damage resistance signature (IRDS) associated with resistance to chemotherapy or radiotherapy in cell lines. This was used to develop a seven gene signature in order to classify patients as IRDSþ (DNA damage resistant) or IRDSe (DNA damage sensitive). The IRDS was initially evaluated in 295 breast cancer patients and was shown to predict metastasis in patients treated with adjuvant chemotherapy versus nonchemotherapy treated patients. IRDSþ patients had higher locoregional failure rates than IRDSe patients after radiotherapy. However, there were too few events to evaluate the effect of IRDS in the non-radiotherapy treated group. In order to confirm its value as a therapy-specific marker, IRDS was then evaluated in three validation cohorts, one treated with adjuvant chemotherapy and/or radiotherapy (cohort A, n ¼ 292), one treated with adjuvant endocrine therapy but no other systemic therapy (cohort B, n ¼ 277) and one group who did not receive adjuvant systemic therapy (cohort C, n ¼ 286). In the first group, IRDSþ patients had shorter RFS than IRDSe patients; IRDS was also associated with metastasis-free survival (MFS) in those receiving adjuvant chemotherapy and with locoregional control in those treated with radiotherapy. There was no association with MFS for the other two cohorts and data regarding whether these groups received radiotherapy were not available. All the radiotherapy treated patients in cohort A received adjuvant chemotherapy. ‘Importance scores’ for the prediction of MFS and locoregional control were calculated for IRDS and other variables and compared in a meta-analysis including all patients. This showed a higher importance score for IRDS in predicting MFS in chemotherapy treated versus nonchemotherapy treated patients and a higher importance score of IRDS in predicting locoregional control in radiotherapy treated versus non-radiotherapy treated patients [33].
DNA Damage Response-related Biomarkers MRE11 MRE11 is a DDR protein that forms part of the MRN complex with RAD50 and NBS-1. MRN acts as a sensor of DSBs, recruits ATM and is involved in end resection during DSB repair [34].
Please cite this article in press as: Forker LJ, et al., Biomarkers of Tumour Radiosensitivity and Predicting Benefit from Radiotherapy, Clinical Oncology (2015), http://dx.doi.org/10.1016/j.clon.2015.06.002
6
L.J. Forker et al. / Clinical Oncology xxx (2015) 1e9
€derlund et al. [35] evaluated the expression of the DDR So proteins ATM, MRE11, RAD50 and NBS1 by immunohistochemistry (IHC) in tumour samples from 224 patients who underwent mastectomy for breast cancer and were randomised between adjuvant chemotherapy and adjuvant radiotherapy. In patients with moderate/strong MRN expression, local recurrence-free survival (LRFS) was higher for patients treated with radiotherapy than for those treated with chemotherapy (P ¼ 0.009), whereas in those with negative/weak MRN expression there was no difference in LRFS between chemotherapy and radiotherapy (P ¼ 0.73). Benefit in terms of local control in radiotherapy compared with chemotherapy was seen in patients with moderate/strong expression of the individual proteins RAD50, NBS1 and MRE11, but not for patients with negative/weak expression [35]. Choudhury et al. [36] assessed expression of the DDR proteins ATM, MRE11, RAD50, NBS1 and H2AX by IHC in two cohorts of muscle invasive bladder cancer (MIBC) patients treated with radical radiotherapy and a single cohort treated with radical cystectomy. In both radical radiotherapy groups, patients with high MRE11 showed improved cause-specific survival (CSS) at 3 years (cohort A high 68.7% versus low 43.1%, P ¼ 0.012; cohort B high 71.2% versus low 43.0%, P ¼ 0.02). However, in patients who underwent radical cystectomy, CSS was similar for high and low MRE11 expression (high 53.8% versus low 62.2%, P ¼ 0.46). There was no relationship with CSS for any of the other potential biomarkers assessed [36]. These results were replicated by Laurberg et al. [37] in a study that assessed expression of p16, ATM, TIP60, Rb, p53, ki67 and MRE11 by IHC in three cohorts of MIBC patients, one treated with radical radiotherapy with concurrent chemotherapy and two managed surgically. In 148 radical radiotherapy patients, those with high MRE11 expression had longer disease-specific survival (P ¼ 0.005). No value in the prediction of disease-specific survival was found for 273 cystectomy patients (P ¼ 0.96). TIP60 was predictive of outcome in cystectomy patients but not radiotherapy patients [37]. Most recently, Teo et al. [38] used next generation sequencing to explore the role of germline variants of MRE11A (the gene encoding MRE11) in MIBC. In 186 patients treated with radical radiotherapy, carriage of at least one of six rare variants was associated with worse 5 year CSS (P ¼ 0.009). The single nucleotide polymorphism (SNP) rs18005363 G > A was also associated with poor 5 year CSS (per allele hazard ratio 2.10, 95% confidence interval 1.34e3.28, P ¼ 0.001); 5 year CSS by genotype was 58.3% GG, 42.0% GA and 0% AA. No association of this SNP with CSS was seen for 256 cystectomy patients (5 year CSS 59.7% GG, 54.0% GA and 75.0% AA, P ¼ 0.89) [38]. AIMP3 AIMP3 is a tumour suppressor gene that acts as an upstream regulator of p53 in response to DNA damage. Gurung et al. [39] showed low AIMP3 expression in MIBC and evaluated AIMP3 and p53 for their ability to predict radiotherapy outcomes. AIMP3 expression was defined as
positive or negative by IHC. In 217 MIBC patients treated with radical radiotherapy carbogen and nicotinamide enrolled in the BCON study, the median survival was superior in AIMP3-positive patients (67.9 months positive versus 21.5 months negative, P ¼ 0.002). In 151 MIBC patients treated with cystectomy, AIMP3 expression did not correlate with survival (P ¼ 0.70). No interaction between AIMP3 expression and either BCON trial arm was found and p53 had no association with survival in radiotherapy or cystectomy cohorts [39]. NBN NBN is the gene encoding NBS-1, which also forms part of the MRN complex [34]. Berlin et al. [40] determined copy number variation in the DDR genes MRE11A, RAD50, NBN, ATM, ATR and PRKDC in pre-treatment prostate cancer biopsies. Variation in NBN copy number was the most common abnormality observed and was associated with genomic instability [40]. In 139 localised prostate cancer patients treated with radical image-guided radiotherapy, NBN gain was a significant predictor of biochemical relapse-free rate at 5 years (46% NBN gain versus 77% NBN neutral, P ¼ 0.00067) and remained significant in a multivariate analysis. In a control cohort of 154 localised prostate cancer patients treated with radical prostatectomy, NBN gain was not associated with biochemical relapse-free rate (61% NBN gain versus 77% NBN neutral, P ¼ 0.15). No association was found for any of the other genes assessed [40]. BRE BRE is a member of the BRCA1 complex, which is important in DSB repair. Noordermeer et al. [41] used quantitative PCR (QPCR) to determine BRE expression in 229 breast cancer patients. For the whole cohort, BRE expression was not prognostic. In 169 patients treated with adjuvant radiotherapy, high BRE expression correlated with improved disease-free survival (P ¼ 0.004) and remained significant on multivariate analysis. High BRE expression initially seemed to predict poor disease-free survival in non-radiotherapy treated patients, but did not remain significant on multivariate analysis. High BRE expression was a favourable prognostic factor in a larger cohort of 2324 breast cancer patients, but could not be assessed for treatment-specific predictive value [41].
Discussion Limitations of Radiotherapy Biomarkers Identified The literature search highlighted that despite many citations identified, few studies have been undertaken in which a radiotherapy biomarker had undergone robust clinical assessment in over 100 patients and was compared between radiotherapy and non-radiotherapy treated groups to confirm its value as a predictive rather than prognostic marker. Ten biomarkers were reported as being predictive of radiotherapy benefit, five were gene signatures and five were DDR-related markers.
Please cite this article in press as: Forker LJ, et al., Biomarkers of Tumour Radiosensitivity and Predicting Benefit from Radiotherapy, Clinical Oncology (2015), http://dx.doi.org/10.1016/j.clon.2015.06.002
L.J. Forker et al. / Clinical Oncology xxx (2015) 1e9
RSI is the most extensively clinically investigated radiosensitivity signature, having been assessed in four different tumour types in five independent cohorts [27,28]. However, the rectal, oesophageal and head and neck cancer datasets represented preliminary work and so included low numbers of patients, findings were not validated in a second cohort or compared with a group of non-radiotherapy treated controls. RSI predicted improved RFS exclusively in patients who received adjuvant radiotherapy in two larger independent cohorts of over 500 breast cancer patients in total. The 31 gene signature investigated in glioma was found to be predictive of longer overall survival in radiotherapy treated patients only in the GSE16011 cohort and in the TCGA cohort in which all patients received radiotherapy [30]. Further work is necessary to conclude that this is definitely a predictive rather than a prognostic marker, as it has only been evaluated in one cohort that allowed comparison between radiotherapy and non-radiotherapy treated patients and was associated with overall survival in both groups on univariate analysis. Neither RSI nor the 31 gene signature has been investigated in a randomised controlled trial (RCT). The other three signatures were investigated for value in predicting benefit from adjuvant radiotherapy in breast cancer. Weichselbaum et al. [33] concluded that the IRDS signature predicted benefit from radiotherapy. However, there is stronger evidence for its value in predicting the response to adjuvant chemotherapy. All of the radiotherapy treated patients in the validation cohort also received chemotherapy and so the poorer locoregional control seen in the IRDSþ group may have been due to chemoresistance rather than radioresistance. It is also unclear as to whether this has truly been assessed in radiotherapy versus nonradiotherapy treated patients, as information was not available regarding radiotherapy for two of the cohorts included in the meta-analysis. MammaPrintÒ is probably a prognostic rather than a predictive biomarker and the Danish Breast Cancer Cooperative Group radiotherapy profile requires further validation in a larger cohort as part of an RCT. The most promising DDR-related biomarker of radiotherapy outcomes identified was MRE11. High tumour MRE11 expression has been shown to correlate with improved LRFS in breast cancer only in patients who received adjuvant radiotherapy [35] and with CSS in MIBC in radical radiotherapy but not cystectomy patients. In MIBC these findings have been validated in a large number of patients; two independent radiotherapy cohorts were included in the first study published [36] and findings have been replicated in another cohort by a different group [37]. It is reasonable to conclude that MRE11 may be a predictive biomarker of benefit from radiotherapy and work is ongoing to assess this in patients previously recruited as part of the phase III RCTs BCON [42] and BC2001 [43]. A germline SNP in MRE11A has also been shown to predict outcome in radiotherapy patients but not cystectomy patients in MIBC, but is yet to be validated in an independent cohort [38]. This highlights a major problem, namely that germline blood
7
samples have not been routinely collected in radiotherapybased studies to date. Other DDR-related biomarkers identified as predictive of outcome exclusively in radiotherapy treated patients were AIMP3 in MIBC [39], NBN in prostate cancer [40] and BRE in breast cancer [41], but these findings have not yet been replicated. Although some of these biomarkers show promise as being predictive of radiotherapy outcomes, particularly RSI and MRE11, none has been assessed in a prospective RCT, which would be the gold standard approach to validate a predictive marker. Therefore, further studies are needed before they can be introduced into standard practice. Challenges in Implementing Biomarker-driven Radiotherapy Despite the growing field of cancer biomarker research and numerous papers published in the literature, no predictive biomarkers of radiotherapy outcome have transitioned to use as the standard of care. One difficulty is in providing levels of evidence high enough to justify routine clinical use. This is particularly difficult for predictive markers, which must be assessed in relation to outcomes in a therapy-specific manner [11,44]. Low patient numbers and lack of a control population were common reasons for exclusion of biomarkers from this review. Predictive biomarker validation requires an RCT in order to ensure the patient groups are comparable and so the collection of samples for translational research is extremely important. Ideally a prospective RCT in which biomarker validation is included in the trial design should be used, but retrospective analysis of previous RCTs is also acceptable [11]. None of the biomarkers identified by this review was compared between patients who received or did not receive radiotherapy in an RCT. Assessment of a radiotherapy biomarker in this way is particularly difficult as RCTs of radical radiotherapy versus non-radiotherapy radical treatments are rare. For example, in MIBC, radical radiotherapy and surgery are said to be comparable in efficacy on the basis of retrospective data [45], but an attempt to evaluate this in an RCT was aborted [46]. If a predictive biomarker were to be integrated into routine practice it would require a convenient, reproducible assay capable of delivering quick results. The REMARK guidelines for reporting of biomarker studies specify the importance of inclusion of a detailed assay method [44,47]. IHC is often used in the assessment of biomarkers in tumour samples, but can give variable results due to inconsistencies in methodology and interpretation [48,49]. The evaluation of biomarkers at the level of RNA may be more easily reproducible. However, for some markers, posttranscriptional modifications mean that mRNA levels and protein levels do not always correlate [50]. SNP genotyping for germline mutations, as used in the study of MRE11A, would be attractive as this could be carried out on a peripheral blood sample. The studies of MRE11 expression used IHC, which would be challenging to implement in routine practice, as this would require rigorous guidelines for standardisation across laboratories, although this has
Please cite this article in press as: Forker LJ, et al., Biomarkers of Tumour Radiosensitivity and Predicting Benefit from Radiotherapy, Clinical Oncology (2015), http://dx.doi.org/10.1016/j.clon.2015.06.002
8
L.J. Forker et al. / Clinical Oncology xxx (2015) 1e9
been achievable for other tests, such as HER-2 in breast cancer [51].
Conclusion Currently there are no biomarkers predictive of radiosensitivity or radiotherapy benefit used as standard of care. There are several on the horizon that show promise but require further clinical assessment. Predictive biomarkers for targeted agents have revolutionised practice in some areas of systemic treatment but these only affect a small proportion of cancer patients. As radiotherapy is such a commonly utilised treatment modality the radiosensitivity biomarkers identified have the potential to improve treatment outcomes for very large numbers of patients and therefore would have a huge effect on oncology practice.
References [1] Ferlay J, Soerjomataram I, Ervik M, et al. GLOBOCAN 2012 v1.0, Cancer Incidence and Mortality Worldwide: IARC CancerBase No. 11 [database on the Internet]. Lyon: International Agency for Research on Cancer. Available at: http://globocan.iarc.fr; 2013. [2] Horwich A, Parker C, de Reijke T, Kataja V. Prostate cancer: ESMO clinical practice guidelines for diagnosis, treatment and follow-up. Ann Oncol 2013;24(Suppl. 6):vi106evi114. [3] Vansteenkiste J, De Ruysscher D, Eberhardt WEE, et al. Early and locally advanced non-small-cell lung cancer (NSCLC): ESMO clinical practice guidelines for diagnosis, treatment and follow-up. Ann Oncol 2013;24(Suppl. 6):vi89evi98. [4] Senkus E, Kyriakides S, Penault-Llorca F, et al. Primary breast cancer: ESMO clinical practice guidelines for diagnosis, treatment and follow-up. Ann Oncol 2013;24(Suppl. 6):vi7evi23. [5] Glimelius B, Tiret E, Cervantes A, Arnold D. Rectal cancer: ESMO clinical practice guidelines for diagnosis, treatment and follow-up. Ann Oncol 2013;24(Suppl. 6):vi81evi88. [6] Lutz ST, Jones J, Chow E. Role of radiation therapy in palliative care of the patient with cancer. J Clin Oncol 2014;32(26): 2913e2919. [7] Delaney G, Jacob S, Featherstone C, Barton M. The role of radiotherapy in cancer treatment. Cancer 2005;104(6): 1129e1137. [8] Staffurth J. A review of the clinical evidence for intensitymodulated radiotherapy. Clin Oncol 2010;22(8):643e657. [9] Hanahan D, Weinberg RA. Hallmarks of cancer: the next generation. Cell 2011;144(5):646e674. [10] Burrell RA, McGranahan N, Bartek J, Swanton C. The causes and consequences of genetic heterogeneity in cancer evolution. Nature 2013;501(7467):338e345. [11] Mandrekar SJ, Sargent DJ. Clinical trial designs for predictive biomarker validation: theoretical considerations and practical challenges. J Clin Oncol 2009;27(24):4027e4034. [12] Polley M-YC, Freidlin B, Korn EL, Conley BA, Abrams JS, McShane LM. Statistical and practical considerations for clinical evaluation of predictive biomarkers. J Natl Cancer Inst 2013;105(22):1677e1683. [13] Slamon DJ, Leyland-Jones B, Shak S, et al. Use of chemotherapy plus a monoclonal antibody against HER2 for metastatic breast cancer that overexpresses HER2. N Engl J Med 2001;344(11):783e792.
s J, Kim S-B, et al. Pertuzumab plus trastuzu[14] Baselga J, Corte mab plus docetaxel for metastatic breast cancer. N Engl J Med 2012;366(2):109e119. [15] Verma S, Miles D, Gianni L, et al. Trastuzumab emtansine for HER2-positive advanced breast cancer. N Engl J Med 2012;367(19):1783e1791. [16] Takano T, Ohe Y, Sakamoto H, et al. Epidermal growth factor receptor gene mutations and increased copy numbers predict gefitinib sensitivity in patients with recurrent nonesmall-cell lung cancer. J Clin Oncol 2005;23(28):6829e6837. [17] Mitsudomi T, Morita S, Yatabe Y, et al. Gefitinib versus cisplatin plus docetaxel in patients with non-small-cell lung cancer harbouring mutations of the epidermal growth factor receptor (WJTOG3405): an open label, randomised phase 3 trial. Lancet Oncol 2010;11(2):121e128. [18] Shaw AT, Yeap BY, Solomon BJ, et al. Effect of crizotinib on overall survival in patients with advanced non-small-cell lung cancer harbouring ALK gene rearrangement: a retrospective analysis. Lancet Oncol 2011;12(11):1004e1012. [19] Solomon BJ, Mok T, Kim D-W, et al. First-line crizotinib versus chemotherapy in ALK-positive lung cancer. N Engl J Med 2014;371(23):2167e2177. [20] West CM, Davidson SE, Roberts SA, Hunter RD. The independence of intrinsic radiosensitivity as a prognostic factor for patient response to radiotherapy of carcinoma of the cervix. Br J Cancer 1997;76(9):1184e1190. [21] van de Vijver MJ, He YD, van’t Veer LJ, et al. A gene-expression signature as a predictor of survival in breast cancer. N Engl J Med 2002;347(25):1999e2009. [22] Paik S, Shak S, Tang G, et al. A multigene assay to predict recurrence of tamoxifen-treated, node-negative breast cancer. N Engl J Med 2004;351(27):817e826. [23] Eustace A, Mani N, Span PN, et al. A 26-gene hypoxia signature predicts benefit from hypoxia-modifying therapy in laryngeal cancer but not bladder cancer. Clin Cancer Res 2013;19(17):4879e4888. [24] Wouters BG, Begg AC. Irradiation-induced damage and the DNA damage response. In: Joiner M, van der Kogel A, editors. Basic clinical radiobiology. Great Britain: CRC Press; 2009. p. 11e26. € hne M, Rief N, et al. A pathway of double-strand [25] Riballo E, Ku break rejoining dependent upon ATM, Artemis and proteins locating to g-H2AX foci. Mol Cell 2004;16(5):715e724. [26] Torres-Roca JF, Eschrich S, Zhao H, et al. Prediction of radiation sensitivity using a gene expression classifier. Cancer Res 2005;65(16):7169e7176. [27] Eschrich SA, Pramana J, Zhang H, et al. A gene expression model of intrinsic tumor radiosensitivity: prediction of response and prognosis after chemoradiation. Int J Radiat Oncol Biol Phys 2009;75(2):489e496. [28] Eschrich SA, Fulp WJ, Pawitan Y, et al. Validation of a radiosensitivity molecular signature in breast cancer. Clin Cancer Res 2012;18(18):5134e5143. [29] Kim H, Kim S, Kim S, et al. Identification of a radiosensitivity signature using integrative metaanalysis of published microarray data for NCI-60 cancer cells. BMC Genomics 2012;13(1):348. [30] Meng J, Li P, Zhang Q, Yang Z, Fu S. A radiosensitivity gene signature in predicting glioma prognostic via EMT pathway. Oncotarget 2014;5(13):4683e4693. [31] Drukker CA, Elias SG, Nijenhuis MV, et al. Gene expression profiling to predict the risk of locoregional recurrence in breast cancer: a pooled analysis. Breast Cancer Res Treat 2014;148(3):599e613. [32] Tramm T, Mohammed H, Myhre S, et al. Development and validation of a gene profile predicting benefit of
Please cite this article in press as: Forker LJ, et al., Biomarkers of Tumour Radiosensitivity and Predicting Benefit from Radiotherapy, Clinical Oncology (2015), http://dx.doi.org/10.1016/j.clon.2015.06.002
L.J. Forker et al. / Clinical Oncology xxx (2015) 1e9
[33]
[34] [35]
[36]
[37]
[38]
[39]
[40]
[41]
postmastectomy radiotherapy in patients with high-risk breast cancer: a study of gene expression in the DBCG82bc cohort. Clin Cancer Res 2014;20(20):5272e5280. Weichselbaum RR, Ishwaran H, Yoon T, et al. An interferonrelated gene signature for DNA damage resistance is a predictive marker for chemotherapy and radiation for breast cancer. Proc Natl Acad Sci USA 2008;105(47):18490e18495. Harper JW, Elledge SJ. The DNA damage response: ten years after. Mol Cell 2007;28(5):739e745. € derlund K, St € ld B, So al O, Skoog L, Rutqvist LE, Nordenskjo Askmalm MS. Intact Mre11/Rad50/Nbs1 complex predicts good response to radiotherapy in early breast cancer. Int J Radiat Oncol Biol Phys 2007;68(1):50e58. Choudhury A, Nelson LD, Teo MTW, et al. MRE11 expression is predictive of cause-specific survival following radical radiotherapy for muscle invasive bladder cancer. Cancer Res 2010;70(18):7017e7026. Laurberg JR, Brems-Eskildsen AS, Nordentoft I, et al. Expression of TIP60 (tat-interactive protein) and MRE11 (meiotic recombination 11 homolog) predict treatment-specific outcome of localised invasive bladder cancer. BJU Int 2012;110(11c):E1228eE1236. Teo MTW, Dyrskjøt L, Nsengimana J, et al. Next-generation sequencing identifies germline MRE11A variants as markers of radiotherapy outcomes in muscle-invasive bladder cancer. Ann Oncol 2014;25(4):877e883. Gurung PMS, Veerakumarasivam A, Williamson M, et al. Loss of expression of the tumour suppressor gene AIMP3 predicts survival following radiotherapy in muscle-invasive bladder cancer. Int J Cancer 2015;136(3):709e720. Berlin A, Lalonde E, Sykes J, et al. NBN gain is predictive for adverse outcome following image-guided radiotherapy for localized prostate cancer. Oncotarget 2014;5(22): 11081e11090. Noordermeer SM, Wennemers M, Bergevoet SM, et al. Expression of the BRCA1 complex member BRE predicts
[42]
[43]
[44]
[45]
[46]
[47]
[48]
[49]
[50]
[51]
9
disease free survival in breast cancer. Breast Cancer Res Treat 2012;135(1):125e133. Hoskin PJ, Rojas AM, Bentzen SM, Saunders MI. Radiotherapy with concurrent carbogen and nicotinamide in bladder carcinoma. J Clin Oncol 2010;28(33):4912e4918. James ND, Hussain SA, Hall E, et al. Radiotherapy with or without chemotherapy in muscle-invasive bladder cancer. N Engl J Med 2012;366(16):1477e1488. Altman DG, McShane LM, Sauerbrei W, Taube SE. Reporting recommendations for tumor marker prognostic studies (REMARK): explanation and elaboration. PLoS Med 2012;9(5):e1001216. Kotwal S, Choudhury A, Johnston C, Paul AB, Whelan P, Kiltie AE. Similar treatment outcomes for radical cystectomy and radical radiotherapy in invasive bladder cancer treated at a United Kingdom specialist treatment center. Int J Radiat Oncol Biol Phys 2008;70(2):456e463. Huddart RA, Hall E, Lewis R, Birtle A, on behalf of the STMG. Life and death of SPARE (Selective Bladder Preservation Against Radical Excision): reflections on why the spare trial closed. BJU Int 2010;106(6):753e755. McShane LM, Altman DG, Sauerbrei W, et al. Reporting recommendations for tumor marker prognostic studies (REMARK). Natl Clin Prac Oncol 2005;2(8):416e422. McShane LM, Aamodt R, Cordon-Cardo C, et al. Reproducibility of p53 immunohistochemistry in bladder tumors. Clin Cancer Res 2000;6(5):1854e1864. Bussolati G, Leonardo E. Technical pitfalls potentially affecting diagnoses in immunohistochemistry. J Clin Pathol 2008; 61(11):1184e1192. Martin RM, Kerr M, Teo MTW, et al. Post-transcriptional regulation of MRE11 expression in muscle-invasive bladder tumours. Oncotarget 2014;5(4):993e1003. Rakha EA, Pinder SE, Bartlett JMS, et al. Updated UK recommendations for HER2 assessment in breast cancer. J Clin Pathol 2014;68(2):93e99.
Please cite this article in press as: Forker LJ, et al., Biomarkers of Tumour Radiosensitivity and Predicting Benefit from Radiotherapy, Clinical Oncology (2015), http://dx.doi.org/10.1016/j.clon.2015.06.002