Spatial Needle Placement Strategies and Pathological Concordance in Prostate Biopsy: A National Multicenter Retrospective Cohort Study

Background: Prostate cancer is the second most common malignancy in men worldwide. GLOBOCAN 2022 estimated about 1.468 million new cases and 397,000 deaths globally, and about 134,000 new cases and 48,000 deaths in China. Biopsy directly affects detection of clinically significant prostate cancer (csPCa) and choices of active surveillance or definitive treatment. Conventional systematic biopsy samples by anatomical zones with a fixed number of cores, yet tumors are multifocal and spatially heterogeneous, so the same total core count may correspond to markedly different coverage of anterior, posterior, apical, basal, peripheral-zone, and transition-zone regions. Although mpMRI and targeted biopsy have entered clinical pathways, and combined targeted plus systematic sampling is common, evidence on which anatomical regions to cover, the extent of perilesional sampling, and how many systematic cores to retain in different risk strata remains limited, especially across China with a unified spatial ontology and core-level quality control. A second challenge is the discordance between biopsy Grade Group and whole-gland radical prostatectomy pathology. Design: This is a national multicenter, retrospective, non-interventional observational cohort study. More than 20 hospitals covering more than 10 provinces will contribute at least 30,000 consecutive audited cases. The lead center's retrospective data window is January 2015 to October 2026; each participating center's window will cover a consecutive, verifiable period specified in its site agreement and ethics materials. Four nested analysis cohorts are defined: (A) all eligible biopsy events, for the national strategy map and observed detection; (B) those with interpretable MRI/PI-RADS data, for MRI risk stratification; (C) those with localizable cores and strategy maps, for within-patient template reconstruction; and (D) those who underwent radical prostatectomy with matched biopsy-surgery pathology, for GG concordance, upgrading, and downgrading. First biopsies form the main cohort; subsequent biopsies of the same patient are recorded as independent events in a prespecified repeat-biopsy cohort. Spatial definitions and reconstruction: Perilesional cores are defined by the shortest distance from a non-targeted needle trajectory to the three-dimensional MRI lesion boundary rather than by zone: within the boundary is targeted; outside the boundary, \<=10 mm is perilesional (subdivided into 0-5 mm and \>5-10 mm), \>10-15 mm is extended-near, and \>15 mm is remote. Within-patient template reconstruction uses all actually obtained, localizable cores as the observable set and simulates retaining targeted-only, targeted plus perilesional, targeted plus ipsilateral systematic, or targeted plus bilateral systematic cores, calculating cores saved, GG\>=2/GG\>=3 missed, and GG1 reduction. This paired analysis does not assume unsampled tissue is negative. Analysis and boundaries: Detection outcomes are defined by biopsy pathology in the full biopsy cohort; whole-gland pathology is used only in the selective surgery subcohort, and absence of surgery is not treated as negative. The study reports observed detection rates, incremental detection, and the absolute omission cost of candidate reduced templates; it does not report traditional sensitivity, specificity, or negative predictive value in the full cohort because no uniform whole-gland reference standard exists. Analyses use center-clustered robust confidence intervals or multilevel models, generalized linear mixed models or marginal standardization for patient-level binary outcomes, nested core-level models, paired risk differences and McNemar tests with patient- and center-clustered bootstrap for within-patient reconstruction, and weighted kappa and confusion matrices for pathology concordance. The study follows STROBE, RECORD, START, and applicable STARD principles. No examinations, biopsies, treatment, or follow-up are added; the main risks are privacy breach and data misuse. Waiver of written informed consent is requested because the study uses only existing clinical data, imposes no intervention or physical risk, and re-contacting a large, geographically dispersed, long-period population is infeasible and would introduce selection bias, subject to de-identification, data minimization, access control, and audit.