Prostate cancer (PCa) is the second most common cancer in men, accounting for 15% of all diagnosed cancers (1). Currently, screening for PCa is one of the most controversial topics in the urological literature. In fact, PCa screening has led to a reduction in advanced disease and disease-specific mortality. However, many cases of overdiagnosis by using prostate-specific antigen (PSA) test were observed, leading to harmful treatments for patient quality of life. Therefore, the primary objectives of screening and early detection are: reduction in mortality and at least, a maintained satisfactory quality of life. Radical treatment, such as surgery and radiotherapy, can negatively impact long-term quality of life causing sexual, urinary, and bowel dysfunction. Moreover, androgen deprivation therapy used in short or longterm treatment can induce different side effects including loss of muscle mass, sexual problems, adverse metabolic sequelae, and increased cardiovascular risk. Many patients with screening-detected localized PCa will not benefit from definitive treatment and 45% of them are candidates for deferred management (active surveillance and watchful waiting) (2). The key point is the need to discriminate “latent” PCa from “lethal” PCa and subsequent over-diagnosis and over-treatment. Therefore, other biomarkers for a more accurate prognosis in PCa are needed. Tissue samples were collected form 48 patients (23 GS6, 11 GS7, 11 GS8 and 3 GS9) with pathologically confirmed PCa, who underwent curative radical prostatectomy, between 2010 and 2012. After 5 years of follow-up, patients were stratified into two groups (Table I) according to their prognosis (benign, n=25; poor, n=23). To confirm Gleason score (GS) and tumor cellularity, all formalin-fixed paraffin-embedded (FFPE) tumor tissue samples were stained with hematoxylin and eosine (H&E) and evaluated by genitourinary pathologist. Selected samples (both tumor and normal tissues from the same patient) were cut into 8×10 μm sections. One 4-μm thick section was stained with H&E to confirm tumor cellularity. Genomic DNA (gDNA) was extracted using the QIAmp FFPE tissue kit (Qiagen, Germantown, MD, USA) according to the manufacturer’s instructions. gDNA was quantified by a Qubit ® 2.0 Fluorometer (Thermo Fisher Scientific, Waltham, MA, USA) using Qubit ® dsDNA HS Assay Kit. Libraries were prepared from 10 ng of gDNA using the Ion AmpliSeq™ OnDemand Panel (PC Panel). The PC Panel was designed using AmpliSeq.com program by selecting target regions of 16 genes (APC, AR, ATM, CDK12, CHD1, COL5A1, FOXA1, MED12, KMT2D, OR5L1, PIK3CA, PTEN, RB1, SPOP, TP53, ZFHX3), which are the more frequently mutated in PCa (cBioPortal database). The panel consisted of two DNA primer pools (pool 1: 337 amplicons, pool 2:331 amplicons) capable to amplify coding regions of maximum 150 bp to ensure optimal amplification. Overall, gDNA was subjected to library preparation according to the Ion Ampliseq Librery kit Plus (Thermo Fisher Scientific, Waltham, MA, USA). Target regions were initially amplified (20 PCR cycles) with a multiple PCR and the amplicons produced from pool 1 and pool 2 were combined and partially digested. Amplicons were then subjected to ligation of barcoded adapters and purified. Before sequencing, libraries were quantified using the Agilent™ 2100 Bioanalyzer™ (Agilent Genomics, Santa Clara, CA, USA) and diluted to 100 pM. Barcoded libraries, combined for maximizing chip use, labor and costs, were subjected to emulsion PCR using OneTouch™ Instrument and enriched by the OneTouch™ ES Instrument using the Ion PGM™ Hi-Q View™ OT2 Kit, following the manufacturer’s instructions. Finally, sequencing was performed on the Ion PGM with the Ion PGM™ Hi-Q View™ Sequencing Kit (Thermo Fisher Scientific, Waltham, MA, USA), loading barcoded samples into a 316 v.2 BD chip (3). Sequencing data analysis was conducted by using Torrent Suite software v. 5.0 (Thermo Fischer Scientific, Milan, Italy). The alignment against a reference genome (hg19) was performed using the Torrent Mapping Alignment Program after low-quality reads removal and adapter sequences trimming. The Variant Caller (VC) plug in was used to identify variations from the reference sequence. To identify pathogenic variations, mutations that did not affect the protein coding regions were filtered out. All identified variants were visually confirmed by the Integrative Genome viewer (IGV). Genomic Evolutionary Rate Profiling (GERP) tools were used to predict the effect of missense mutations on the protein and calculate their conservation scores. All 48 tumor and matched normal samples were sequenced. All target genes of the PC Panel were covered and the minimum coverage was 500×. Despite this high coverage, 5/48 FFPE samples revealed at least 50 amplicons with coverage between 500× and 100× and 20 amplicons with a coverage lower than 100×, suggesting that lower quality of gDNA could affect sequencing results. The PC Panel design had a high performance with high copy number of all amplicons except for the AMPL 7153036487 region, encoding AR gene (start 66765084-end 66765219, 136 bp), which was missing from the sequencing in all samples. The VC plug-in reported a total of about 3000 mutations (80% single nucleotide variants, SNVs and 20% indels), but they were successively filtered for coverage and mutation frequency. In fact, all variants with coverage lower than 100× and mutation frequency less than 8-10% were not considered. Moreover, results from normal and tumor tissue were compared and variants present in both samples were excluded. Finally, a total of 95 variants were selected for annotation (92 SNVs, 2 indels, and 1 duplication). The 95 variants were found in 14 genes of the PC Panel, while no variants were annotated in two genes (OR5L1, CDH1). Analyzing the variant distribution between the selected genes, we found that KMT2D, AR, ATM, TP53, FOXA1, and CDK12 are the most frequently mutated genes in more than 50% of our cohort. MED12, ZFHX3, SPOP, and APC genes showed variants in about 30% of tumor samples, while the lower mutation frequency was observed in PTEN, RB1, COL5A1 and PIK3CA genes (Table II). Moreover, by matching gene mutations with the follow-up of patients, we observed that the gene variant frequency is different between patients with benign or poor prognosis. In fact, the mutation frequency is increased in ATM, ZFHX3, SPOP, APC, RB1, and TP53 in patients with poor prognosis, while variants present in KMT2D, COL5A1, AR, and CDK12 are decreased in the same population (Table III). No substantial variations were shown for the rest of the genes. NGS analysis is a powerful approach to detect genomic lesions in PCa starting from low amount of DNA also in critical condition (FFPE Tissues). Our results are in line with data reported in the comprehensive databases published online indicating that PC Panel could individuate tumor variants that are involved in PCa in an efficiently manner. Moreover, some of the annotated variants were already known as recurrent in PCa and they were studied for their clinical implication in initiation, progression and pharmacological impact of PCa. Data from the two groups of patients indicated that “PC Panel” could individuate genome variations typical of PCa progression also in early disease stages. Importantly, somatic mutation in ATM, SPOP, TP53, and KMT2D, which are associated with PCa progression, may be also identified in patients with low Gleason score. In fact, it is known that PTEN and TP53 lesions are more frequently mutated in patients with advanced disease and, likely, the co-localization with SPOP mutations is associated with poor prognosis (4). Moreover, mutation in SPOP gene may promote a more aggressive clinical behavior by increasing genome instability. In conclusion, targeted sequencing approach could increase the possibility to distinguish the patient risk profile improving PCa management strategies.