MP35-11 METABOLIC BIOMARKERS ARE PREDICTIVE OF RESPONSE TO RADIATION THERAPY IN PROSTATE CANCER PATIENTS

You have accessJournal of UrologyProstate Cancer: Markers I1 Apr 2018MP35-11 METABOLIC BIOMARKERS ARE PREDICTIVE OF RESPONSE TO RADIATION THERAPY IN PROSTATE CANCER PATIENTS Amrita Cheema, Scott Grindrod, Simeng Suy, Xiaogang Zhong, Shreyans Jain, Khyati Mehta, Gaurav Bandi, Keith Kowalczyk, John Lynch, Sean Collins, and Anatoly Dritschilo Amrita CheemaAmrita Cheema More articles by this author , Scott GrindrodScott Grindrod More articles by this author , Simeng SuySimeng Suy More articles by this author , Xiaogang ZhongXiaogang Zhong More articles by this author , Shreyans JainShreyans Jain More articles by this author , Khyati MehtaKhyati Mehta More articles by this author , Gaurav BandiGaurav Bandi More articles by this author , Keith KowalczykKeith Kowalczyk More articles by this author , John LynchJohn Lynch More articles by this author , Sean CollinsSean Collins More articles by this author , and Anatoly DritschiloAnatoly Dritschilo More articles by this author View All Author Informationhttps://doi.org/10.1016/j.juro.2018.02.1124AboutPDF ToolsAdd to favoritesDownload CitationsTrack CitationsPermissionsReprints ShareFacebookTwitterLinked InEmail INTRODUCTION AND OBJECTIVES Radiation therapy (RT) is an effective modality as a primary treatment of cancers, or as an adjuvant to surgery or chemotherapy. Risks, benefits and late effects of radiation therapy are observed in the heterogeneous clinical responses of patients receiving curative radiation therapy for prostate cancer. However, about 10 % of treated patients experience radiation treatment related late effects that adversely affect quality of life. The manifestation of these symptoms takes months to develop and raises an urgent need for developing smarter strategies for symptom anticipation and management. Currently there is no blood test that can be used to predict or monitor adverse symptoms in sensitive sub-populations of patients receiving radiation therapy. We tested the hypothesis that prostate cancer patients susceptible to radiation induced adverse effects carry a biochemical fingerprint that could be characterized using blood based metabolomics. Based on a retrospective clinical outcome study, we were able to delineate biomarker panels predictive of radiation responses in patients treated for prostate cancer. METHODS We performed multiple reaction monitoring based targeted metabolomic/lipidomic analyses using baseline (samples obtained before radiation therapy) from the plasma of a cohort of 100 prostate cancer patients. Pre-processed data (using Targetlynx v 3.0) along with clinical annotations were used for the analyses. An optimized classifier algorithm was developed, using baseline plasma samples collected prior to the initiation of radiation therapy. In the retrospective outcome analysis study design, we compared groups of sub-sets of patients who developed adverse symptoms to those who did not (control group), over the course of a two year follow-up interval. Feature selection for building classifiers was performed using LASSO Logistic regression to avoid over-fitting of the model through iterative selection, and cross validation of candidate markers. The top ranking markers were used to create ROC curves to describe the performance of identified biomarkers as diagnostic tests measured on a continuous scale. RESULTS Leveraging clinical outcome data, we performed retrospective outcome analysis to predict radiation toxicities (rectal bleeding, urinary flare and tumor recurrence) by comparing metabolic profiles at baseline. We hypothesized that a specific plasma metabolic bio-signature, prior to radiation therapy, may characterize clinical susceptibility. Metabolite signatures predictive of adverse responses to radiation therapy were developed in the patient cohort treated with stereotactic body radiation therapy (SBRT) for prostate cancer. Twenty nine patients developed transient, symptomatic urinary flare (USF) or obstructed voiding symptoms (UR) and twenty two patients experienced radiation proctitis (RP) or vascular changes on proctoscopy. We were able to develop high accuracy, predictive algorithms for recurrence, urinary symptoms and rectal proctitis episodes in this cohort. The metabolite panels help predict late effects of radiation therapy, and lay the foundation for the development of strategies by which toxicity may be detected at an early stage and mitigated with intervention therapies. CONCLUSIONS We developed candidate predictive biomarker panels that can be used for identifying patients who are likely to develop adverse symptoms following radiation therapy. Such biomarker panels may aid in early detection of tissue toxicity in cancer patients, informing clinical decisions for treatment and follow-up management in patients at risk. © 2018FiguresReferencesRelatedDetails Volume 199Issue 4SApril 2018Page: e451 Advertisement Copyright & Permissions© 2018MetricsAuthor Information Amrita Cheema More articles by this author Scott Grindrod More articles by this author Simeng Suy More articles by this author Xiaogang Zhong More articles by this author Shreyans Jain More articles by this author Khyati Mehta More articles by this author Gaurav Bandi More articles by this author Keith Kowalczyk More articles by this author John Lynch More articles by this author Sean Collins More articles by this author Anatoly Dritschilo More articles by this author Expand All Advertisement Advertisement PDF downloadLoading ...