The REQUITE project: validating predictive models and biomarkers of radiotherapy toxicity to reduce side-effects and improve quality of life in cancer survivors

Radiogenomics and radiotherapy response modeling

Advances in patient-specific information and biotechnology have contributed to a new era of computational medicine. Radiogenomics has emerged as a new field that investigates the role of genetics in treatment response to radiation therapy. Radiation oncology is currently attempting to embrace these recent advances and add to its rich history by maintaining its prominent role as a quantitative leader in oncologic response modeling. Here, we provide an overview of radiogenomics starting with genotyping, data aggregation, and application of different modeling approaches based on modifying traditional radiobiological methods or application of advanced machine learning techniques. We highlight the current status and potential for this new field to reshape the landscape of outcome modeling in radiotherapy and drive future advances in computational oncology.

Symptom clusters for revising scale membership in the analysis of prostate cancer patient reported outcome measures: a secondary data analysis of the Medical Research Council RT01 trial (ISCRTN47772397)

PURPOSE

To investigate the role of symptom clusters in the analysis and utilisation of patient reported outcome measures (PROMs) for data modelling and clinical practice. To compare symptom clusters with scales, and to explore their value in PROMs interpretation and symptom management.

METHODS

A dataset called RT01 (ISCRTN47772397) of 843 prostate cancer patients was used. PROMs were reported with the University of California, Los Angeles Prostate Cancer Index (UCLA-PCI). Symptom clusters were explored with hierarchical cluster analysis (HCA) and average linkage method (correlation > 0.6). The reliability of the Urinary Function Scale was evaluated with Cronbach's Alpha. The strength of the relationship between the items was investigated with Spearman's correlation. Predictive accuracy of the clusters was compared to the scales by receiver operating characteristic (ROC) analysis. Presence of urinary symptoms at 3 years measured with the late effects on normal tissue: subjective, objective, management tool (LENT/SOM) was an endpoint.

RESULTS

Two symptom clusters were identified (urinary cluster and sexual cluster). The grouping of symptom clusters was different than UCLA-PCI Scales. Two items of the urinary function scales ("number of pads" and "urinary leak interfering with sex") were excluded from the urinary cluster. The correlation with the other items in the scale ranged from 0.20 to 0.21 and 0.31 to 0.39, respectively. Cronbach's Alpha showed low correlation of those items with the Urinary Function Scale (0.14-0.36 and 0.33-0.44, respectively). All urinary function scale items were subject to a ceiling effect. Clusters had better predictive accuracy, AUC = 0.70 -0.65, while scales AUC = 0.67-0.61.

CONCLUSION

This study adds to the knowledge on how cluster analysis can be applied for the interpretation and utilisation of PROMs. We conclude that multiple-item scales should be evaluated and that symptom clusters provide a study-specific approach for modelling and interpretation of PROMs.

Data-Based Radiation Oncology: Design of Clinical Trials in the Toxicity Biomarkers Era

The ability to stratify patients using a set of biomarkers, which predict that toxicity risk would allow for radiotherapy (RT) modulation and serve as a valuable tool for precision medicine and personalized RT. For patients presenting with tumors with a low risk of recurrence, modifying RT schedules to avoid toxicity would be clinically advantageous. Indeed, for the patient at low risk of developing radiation-associated toxicity, use of a hypofractionated protocol could be proposed leading to treatment time reduction and a cost-utility advantage. Conversely, for patients predicted to be at high risk for toxicity, either a more conformal form or a new technique of RT, or a multidisciplinary approach employing surgery could be included in the trial design to avoid or mitigate RT when the potential toxicity risk may be higher than the risk of disease recurrence. In addition, for patients at high risk of recurrence and low risk of toxicity, dose escalation, such as a greater boost dose, or irradiation field extensions could be considered to improve local control without severe toxicities, providing enhanced clinical benefit. In cases of high risk of toxicity, tumor control should be prioritized. In this review, toxicity biomarkers with sufficient evidence for clinical testing are presented. In addition, clinical trial designs and predictive models are described for different clinical situations.

Individualized early death and long-term survival prediction after stereotactic radiosurgery for brain metastases of non-small cell lung cancer: Two externally validated nomograms

INTRODUCTION

Commonly used clinical models for survival prediction after stereotactic radiosurgery (SRS) for brain metastases (BMs) are limited by the lack of individual risk scores and disproportionate prognostic groups. In this study, two nomograms were developed to overcome these limitations.

METHODS

495 patients with BMs of NSCLC treated with SRS for a limited number of BMs in four Dutch radiation oncology centers were identified and divided in a training cohort (n=214, patients treated in one hospital) and an external validation cohort n=281, patients treated in three other hospitals). Using the training cohort, nomograms were developed for prediction of early death (<3months) and long-term survival (>12months) with prognostic factors for survival. Accuracy of prediction was defined as the area under the curve (AUC) by receiver operating characteristics analysis for prediction of early death and long term survival. The accuracy of the nomograms was also tested in the external validation cohort.

RESULTS

Prognostic factors for survival were: WHO performance status, presence of extracranial metastases, age, GTV largest BM, and gender. Number of brain metastases and primary tumor control were not prognostic factors for survival. In the external validation cohort, the nomogram predicted early death statistically significantly better (p<0.05) than the unfavorable groups of the RPA, DS-GPA, GGS, SIR, and Rades 2015 (AUC=0.70 versus range AUCs=0.51-0.60 respectively). With an AUC of 0.67, the other nomogram predicted 1year survival statistically significantly better (p<0.05) than the favorable groups of four models (range AUCs=0.57-0.61), except for the SIR (AUC=0.64, p=0.34). The models are available on www.predictcancer.org.

CONCLUSION

The nomograms predicted early death and long-term survival more accurately than commonly used prognostic scores after SRS for a limited number of BMs of NSCLC. Moreover these nomograms enable individualized probability assessment and are easy into use in routine clinical practice.

Radiogenomics: A systems biology approach to understanding genetic risk factors for radiotherapy toxicity?

Adverse reactions in normal tissue after radiotherapy (RT) limit the dose that can be given to tumour cells. Since 80% of individual variation in clinical response is estimated to be caused by patient-related factors, identifying these factors might allow prediction of patients with increased risk of developing severe reactions. While inactivation of cell renewal is considered a major cause of toxicity in early-reacting normal tissues, complex interactions involving multiple cell types, cytokines, and hypoxia seem important for late reactions. Here, we review 'omics' approaches such as screening of genetic polymorphisms or gene expression analysis, and assess the potential of epigenetic factors, posttranslational modification, signal transduction, and metabolism. Furthermore, functional assays have suggested possible associations with clinical risk of adverse reaction. Pathway analysis incorporating different 'omics' approaches may be more efficient in identifying critical pathways than pathway analysis based on single 'omics' data sets. Integrating these pathways with functional assays may be powerful in identifying multiple subgroups of RT patients characterised by different mechanisms. Thus 'omics' and functional approaches may synergise if they are integrated into radiogenomics 'systems biology' to facilitate the goal of individualised radiotherapy.

Pitfalls in Prediction Modeling for Normal Tissue Toxicity in Radiation Therapy: An Illustration With the Individual Radiation Sensitivity and Mammary Carcinoma Risk Factor Investigation Cohorts

PURPOSE

To identify the main causes underlying the failure of prediction models for radiation therapy toxicity to replicate.

METHODS AND MATERIALS

Data were used from two German cohorts, Individual Radiation Sensitivity (ISE) (n=418) and Mammary Carcinoma Risk Factor Investigation (MARIE) (n=409), of breast cancer patients with similar characteristics and radiation therapy treatments. The toxicity endpoint chosen was telangiectasia. The LASSO (least absolute shrinkage and selection operator) logistic regression method was used to build a predictive model for a dichotomized endpoint (Radiation Therapy Oncology Group/European Organization for the Research and Treatment of Cancer score 0, 1, or ≥2). Internal areas under the receiver operating characteristic curve (inAUCs) were calculated by a naïve approach whereby the training data (ISE) were also used for calculating the AUC. Cross-validation was also applied to calculate the AUC within the same cohort, a second type of inAUC. Internal AUCs from cross-validation were calculated within ISE and MARIE separately. Models trained on one dataset (ISE) were applied to a test dataset (MARIE) and AUCs calculated (exAUCs).

RESULTS

Internal AUCs from the naïve approach were generally larger than inAUCs from cross-validation owing to overfitting the training data. Internal AUCs from cross-validation were also generally larger than the exAUCs, reflecting heterogeneity in the predictors between cohorts. The best models with largest inAUCs from cross-validation within both cohorts had a number of common predictors: hypertension, normalized total boost, and presence of estrogen receptors. Surprisingly, the effect (coefficient in the prediction model) of hypertension on telangiectasia incidence was positive in ISE and negative in MARIE. Other predictors were also not common between the 2 cohorts, illustrating that overcoming overfitting does not solve the problem of replication failure of prediction models completely.

CONCLUSIONS

Overfitting and cohort heterogeneity are the 2 main causes of replication failure of prediction models across cohorts. Cross-validation and similar techniques (eg, bootstrapping) cope with overfitting, but the development of validated predictive models for radiation therapy toxicity requires strategies that deal with cohort heterogeneity.

How Will Big Data Improve Clinical and Basic Research in Radiation Therapy?

Historically, basic scientists and clinical researchers have transduced reality into data so that they might explain or predict the world. Because data are fundamental to their craft, these investigators have been on the front lines of the Big Data deluge in recent years. Radiotherapy data are complex and longitudinal data sets are frequently collected to track both tumor and normal tissue response to therapy. As basic, translational and clinical investigators explore with increasingly greater depth the complexity of underlying disease processes and treatment outcomes, larger sample populations are required for research studies and greater quantities of data are being generated. In addition, well-curated research and trial data are being pooled in public data repositories to support large-scale analyses. Thus, the tremendous quantity of information produced in both basic and clinical research in radiation therapy can now be considered as having entered the realm of Big Data.

Radiation-induced CD8 T-lymphocyte Apoptosis as a Predictor of Breast Fibrosis After Radiotherapy: Results of the Prospective Multicenter French Trial

BACKGROUND

Monocentric cohorts suggested that radiation-induced CD8 T-lymphocyte apoptosis (RILA) can predict late toxicity after curative intent radiotherapy (RT). We assessed the role of RILA as a predictor of breast fibrosis (bf +) after adjuvant breast RT in a prospective multicenter trial.

METHODS

A total of 502 breast-cancer patients (pts) treated by conservative surgery and adjuvant RT were recruited at ten centers. RILA was assessed before RT by flow cytometry. Impact of RILA on bf + (primary endpoint) or relapse was assessed using a competing risk method. Receiver-operator characteristic (ROC) curve analyses were also performed in intention to treat. This study is registered with ClinicalTrials.gov, number NCT00893035 and final analyses are presented here.

FINDINGS

Four hundred and fifty-six pts (90.8%) were included in the final analysis. One hundred and eight pts (23.7%) received whole breast and node irradiation. A boost dose of 10-16 Gy was delivered in 449 pts (98.5%). Adjuvant hormonotherapy was administered to 349 pts (76.5%). With a median follow-up of 38.6 months, grade ≥ 2 bf + was observed in 64 pts (14%). A decreased incidence of grade ≥ 2 bf + was observed for increasing values of RILA (p = 0.012). No grade 3 bf + was observed for patients with RILA ≥ 12%. The area under the ROC curve was 0.62. For cut-off values of RILA ≥ 20% and < 12%, sensitivity and specificity were 80% and 34%, 56% and 67%, respectively. Negative predictive value for grade ≥ 2 bf + was equal to 91% for RILA ≥ 20% and positive predictive value was equal to 22% for RILA < 12% where the overall prevalence of grade ≥ 2 bf + was estimated at 14%. A significant decrease in the risk of grade ≥ 2 bf + was found if patients had no adjuvant hormonotherapy (sHR = 0.31, p = 0.007) and presented a RILA ≥ 12% (sHR = 0.45, p = 0.002).

INTERPRETATION

RILA significantly predicts the risk of breast fibrosis. This study validates the use of RILA as a rapid screening test before RT delivery and will change definitely our daily clinical practice in radiation oncology.

FUNDING

The French National Cancer Institute (INCa) through the "Program Hospitalier de Recherche Clinique (PHRC)".

Chronic radiation proctitis: tricks to prevent and treat

OBJECTIVE

The purpose of this study was to give an overview of the measures used to prevent chronic radiation proctitis (CRP) and to provide an algorithm for the treatment of CRP.

METHODS

Medical literature databases including PubMed and Medline were screened and critically analyzed for relevance in the scope of our purpose.

RESULTS

CRP is a relatively frequent late side effect (5-20%) and mainly dependent on the dose and volume of irradiated rectum. Radiation treatment (RT) techniques to prevent CRP are constantly improving thanks to image-guided RT and intensity-modulated RT. Also, newer techniques like protons and new devices such as rectum spacers and balloons have been developed to spare rectal structures. Biopsies do not contribute to diagnosing CRP and should be avoided because of the risk of severe rectal wall damage, such as necrosis and fistulas. There is no consensus on the optimal treatment of CRP. A variety of possibilities is available and includes topical and oral agents, hyperbaric oxygen therapy, and endoscopic interventions.

CONCLUSIONS

CRP has a natural history of improving over time, even without treatment. This is important to take into account when considering these treatments: first be conservative (topical and oral agents) and be aware that invasive treatments can be very toxic.

XRCC1 Polymorphism Associated With Late Toxicity After Radiation Therapy in Breast Cancer Patients

PURPOSE

To identify single-nucleotide polymorphisms (SNPs) in oxidative stress-related genes associated with risk of late toxicities in breast cancer patients receiving radiation therapy.

METHODS AND MATERIALS

Using a 2-stage design, 305 SNPs in 59 candidate genes were investigated in the discovery phase in 753 breast cancer patients from 2 prospective cohorts from Germany. The 10 most promising SNPs in 4 genes were evaluated in the replication phase in up to 1883 breast cancer patients from 6 cohorts identified through the Radiogenomics Consortium. Outcomes of interest were late skin toxicity and fibrosis of the breast, as well as an overall toxicity score (Standardized Total Average Toxicity). Multivariable logistic and linear regression models were used to assess associations between SNPs and late toxicity. A meta-analysis approach was used to summarize evidence.

RESULTS

The association of a genetic variant in the base excision repair gene XRCC1, rs2682585, with normal tissue late radiation toxicity was replicated in all tested studies. In the combined analysis of discovery and replication cohorts, carrying the rare allele was associated with a significantly lower risk of skin toxicities (multivariate odds ratio 0.77, 95% confidence interval 0.61-0.96, P=.02) and a decrease in Standardized Total Average Toxicity scores (-0.08, 95% confidence interval -0.15 to -0.02, P=.016).

CONCLUSIONS

Using a stage design with replication, we identified a variant allele in the base excision repair gene XRCC1 that could be used in combination with additional variants for developing a test to predict late toxicities after radiation therapy in breast cancer patients.

Exploiting biological and physical determinants of radiotherapy toxicity to individualize treatment

The recent advances in radiation delivery can improve tumour control probability (TCP) and reduce treatment-related toxicity. The use of intensity-modulated radiotherapy (IMRT) in particular can reduce normal tissue toxicity, an objective in its own right, and can allow safe dose escalation in selected cases. Ideally, IMRT should be combined with image guidance to verify the position of the target, since patients, target and organs at risk can move day to day. Daily image guidance scans can be used to identify the position of normal tissue structures and potentially to compute the daily delivered dose. Fundamentally, it is still the tolerance of the normal tissues that limits radiotherapy (RT) dose and therefore tumour control. However, the dose-response relationships for both tumour and normal tissues are relatively steep, meaning that small dose differences can translate into clinically relevant improvements. Differences exist between individuals in the severity of toxicity experienced for a given dose of RT. Some of this difference may be the result of differences between the planned dose and the accumulated dose (DA). However, some may be owing to intrinsic differences in radiosensitivity of the normal tissues between individuals. This field has been developing rapidly, with the demonstration of definite associations between genetic polymorphisms and variation in toxicity recently described. It might be possible to identify more resistant patients who would be suitable for dose escalation, as well as more sensitive patients for whom toxicity could be reduced or avoided. Daily differences in delivered dose have been investigated within the VoxTox research programme, using the rectum as an example organ at risk. In patients with prostate cancer receiving curative RT, considerable daily variation in rectal position and dose can be demonstrated, although the median position matches the planning scan well. Overall, in 10 patients, the mean difference between planned and accumulated rectal equivalent uniform doses was -2.7 Gy (5%), and a dose reduction was seen in 7 of the 10 cases. If dose escalation was performed to take rectal dose back to the planned level, this should increase the mean TCP (as biochemical progression-free survival) by 5%. Combining radiogenomics with individual estimates of DA might identify almost half of patients undergoing radical RT who might benefit from either dose escalation, suggesting improved tumour cure or reduced toxicity or both.

Personalized radiotherapy: concepts, biomarkers and trial design

In the past decade, and pointing onwards to the immediate future, clinical radiotherapy has undergone considerable developments, essentially including technological advances to sculpt radiation delivery, the demonstration of the benefit of adding concomitant cytotoxic agents to radiotherapy for a range of tumour types and, intriguingly, the increasing integration of targeted therapeutics for biological optimization of radiation effects. Recent molecular and imaging insights into radiobiology will provide a unique opportunity for rational patient treatment, enabling the parallel design of next-generation trials that formally examine the therapeutic outcome of adding targeted drugs to radiation, together with the critically important assessment of radiation volume and dose-limiting treatment toxicities. In considering the use of systemic agents with presumed radiosensitizing activity, this may also include the identification of molecular, metabolic and imaging markers of treatment response and tolerability, and will need particular attention on patient eligibility. In addition to providing an overview of clinical biomarker studies relevant for personalized radiotherapy, this communication will highlight principles in addressing clinical evaluation of combined-modality-targeted therapeutics and radiation. The increasing number of translational studies that bridge large-scale omics sciences with quality-assured phenomics end points-given the imperative development of open-source data repositories to allow investigators the access to the complex data sets-will enable radiation oncology to continue to position itself with the highest level of evidence within existing clinical practice.

Radiogenomics helps to achieve personalized therapy by evaluating patient responses to radiation treatment

Radiogenomics is the whole genome application of radiogenetics, which focuses on uncovering the underlying genetic causes of individual variation in sensitivity to radiation. There is a growing consensus that radiosensitivity is a complex, inherited polygenic trait, dependent on the interaction of many genes involved in multiple cell processes. An understanding of the genes involved in processes such as DNA damage response and oxidative stress response, has evolved toward examination of how genetic variants, most often, single nucleotide polymorphisms (SNPs), may influence interindividual radioresponse. Many experimental approaches, such as candidate SNP association studies, genome-wide association studies and massively parallel sequencing are being proposed to address these questions. We present a review focusing on recent advances in association studies of SNPs to radiotherapy response and discuss challenges and opportunities for further studies. We also highlight the clinical perspective of radiogenomics in the future of personalized treatment in radiation oncology.