Metabolomic Prediction of Human Prostate Cancer Aggressiveness: Magnetic Resonance Spectroscopy of Histologically Benign Tissue

Abstract

Prostate cancer alters cellular metabolism through events potentially preceding cancer morphological formation. Magnetic resonance spectroscopy (MRS)-based metabolomics of histologically-benign tissues from cancerous prostates can predict disease aggressiveness, offering clinically-translatable prognostic information. This retrospective study of 185 patients (2002-2009) included prostate tissues from prostatectomies (n = 365), benign prostatic hyperplasia (BPH) (n = 15), and biopsy cores from cancer-negative patients (n = 14). Tissues were measured with high resolution magic angle spinning (HRMAS) MRS, followed by quantitative histology using the Prognostic Grade Group (PGG) system. Metabolic profiles, measured solely from 338 of 365 histologically-benign tissues from cancerous prostates and divided into training-testing cohorts, could identify tumor grade and stage, and predict recurrence. Specifically, metabolic profiles: (1) show elevated myo-inositol, an endogenous tumor suppressor and potential mechanistic therapy target, in patients with highly-aggressive cancer, (2) identify a patient sub-group with less aggressive prostate cancer to avoid overtreatment if analysed at biopsy; and (3) subdivide the clinicopathologically indivisible PGG2 group into two distinct KaplanMeier recurrence groups, thereby identifying patients more at-risk for recurrence. Such findings, achievable by biopsy or prostatectomy tissue measurement, could inform treatment strategies. Metabolomics information can help transform a morphology-based diagnostic system by invoking cancer biology to improve evaluation of histologically-benign tissues in cancer environments.The utility of the serum prostate specific antigen (PSA) screening test is now well established, both by its capacity to reveal early-stage disease and its discovery of almost all newly-detected prostate cancers 1-3 . However, despite being prostate-specific, PSA testing is not cancer-specific. Current medical imaging detects 44-87% 4 of clinically significant cancers but remains challenged by small lesions. For this reason, prostate cancer (PCa) diagnoses are still only positively confirmed by prostate transrectal biopsies, performed after an elevated PSA result. After a positive biopsy, an initial cancer grade is assigned according to histological analysis of the biopsy cores; this grade will help determine whether the prostate should be surgically removed by prostatectomy or the patient should enter active surveillance. If a prostatectomy is performed, histological examination of the entire prostate will determine the pathological stage of the disease. With this information, follow-up therapies can be recommended, if warranted.Statistics on prostate cancer's natural history suggest that >70% of patients diagnosed after PSA screening are likely to experience an indolent disease course that little impacts their well-being. About 17% of these newly PSA-diagnosed patients, however, will confront an aggressive prostate cancer that significantly impairs function

Metabolomic prostate cancer fields in HRMAS MRS‐profiled histologically benign tissue vary with cancer status and distance from cancer

“…By considering the ensemble of all measurable low molecular weight metabolites in a biological system, many ex vivo metabolomic studies have indicated the greater sensitivity of global metabolite profiles, with or without univariate metabolite analysis, in representing disease status compared with single metabolites . While ex vivo metabolomic evaluations can assist current clinical diagnosis and staging by providing biological information that cannot be obtained by histological evaluation, through the existence of cancer field effects, translating ex vivo findings to in vivo methods is the ultimate aim. A current challenge to this translation is that in vivo MRS studies typically focus on a select few, high‐intensity metabolites, primarily due to resolution constraints.…”

“…Our review of the literature indicated that investigations of human cancer tissue with HRMAS MRS‐based metabolomics have been conducted in many organ systems since 2004, including adrenal, bone, brain, breast, colorectal, esophagus, lung, pancreas, prostate, rectum, skin, stomach, thyroid, etc. Selected papers with sample sizes of 100 or more are summarized in Table .…”

“…The concept of cancer fields was first used by Dr Danely Slaughter and co‐workers in 1953 to describe histological abnormalities in grossly normal‐appearing tissue and explain multifocality and local recurrence in oral squamous cell carcinoma, and is deemed one of the landmark concepts of cancer research in the past 100 years . However, it has now evolved to describe disease‐related molecular alterations in microscopically normal or histologically benign tissue, observed in cancers of many organ systems and detectable at the epigenomic, genomic, proteomic, or metabolomic levels . Metabolomic field effects were reported in a APCMin/+ mouse model with gastrointestinal tumorigenesis and in human esophageal cancer, oral squamous cell carcinoma, colorectal cancer and prostate cancer .…”

“…However, it has now evolved to describe disease‐related molecular alterations in microscopically normal or histologically benign tissue, observed in cancers of many organ systems and detectable at the epigenomic, genomic, proteomic, or metabolomic levels . Metabolomic field effects were reported in a APCMin/+ mouse model with gastrointestinal tumorigenesis and in human esophageal cancer, oral squamous cell carcinoma, colorectal cancer and prostate cancer . Most significantly, and of great clinical translational potential, histologically benign tissue adjacent to human colorectal cancer could not only predict 5‐year survival with high accuracy (area under the curve, AUC = 0.88) but distinguish tumor stage with higher predictive capacity than results from metabolomic measurements from tumor tissue itself .…”

“…In histologically benign tissue samples a GS of 3 + 3 versus 3 + 4 in adjacent cancer at various distances was strongly correlated with myo‐inositol/scyllo‐inositol ( p = 0.002) and choline/creatine ( p < 0.001) ratios. Most recently, we have reported that analyses of histologically benign prostate tissue from cancerous prostates allow differentiations of not only GS but also pathological stage, and recurrence potential of human prostate cancer . We attributed these distinctions to the existence of metabolomic fields.…”