Prostate Biopsy Pathology

What it does

Prostate Biopsy Pathology reads a whole needle-biopsy case. A systematic biopsy often yields 10 to 12 or more cores, each scanned as an H&E whole-slide image; the run takes every core of the case together, so the result answers the questions of the sign-out: which cores carry carcinoma, how much, and of what grade.

For each core the result gives a carcinoma probability with a heatmap, the tumour length in millimetres and the percentage of the core involved. Cancer-bearing cores receive a Gleason pattern map, Gleason score, ISUP grade group 1–5, the percentage of pattern 4 and 5, and a cribriform flag. Core results then roll up into a case-level summary ordered by part. Prostatectomy slides are accepted too. A biochemical-recurrence risk score with a time-to-event estimate completes the result.

Intended use

Decision support and a second read before sign-out: catching a small focus in a long core, checking a Gleason 3+4 against 4+3, or quantifying pattern 4 for an active-surveillance decision. Benign mimickers (atrophy, adenosis, high-grade PIN, seminal vesicle), treated prostate after radiation or hormonal therapy, and crushed or thin cores are the known pitfalls of the read, and quality flags mark crushed or thin cores.

Who it is for

General surgical pathologists without uropathology fellowship training, pathology residents, and small labs that lack a second reader. Developers call the same run to add prostate grading to a slide-management system or LIS.

Accepted input

• Slides: one H&E whole-slide image per core or part, as DICOM VL Whole Slide Microscopy Image or a vendor file (SVS, NDPI, MRXS, iSyntax, SCN, BIF, CZI, TIFF/OME-TIFF).
• Case: upload all slides of the case at once, each with its part label (for example "left base") so core findings map back to the biopsy site.
• Stain: H&E only. PIN-4 or basal-cell IHC slides are not read by this model.

Requirements

Requirement	Why
Scan at 20× or 40×	Gland architecture for Gleason patterns is read at this magnification
Pixel spacing in µm/px (about 0.5 at 20×, 0.25 at 40×)	Tumour length in mm comes from calibrated spacing
Stain recorded in the slide metadata or the upload	Routes the slide to the H&E read
Focused, unfolded tissue	Blur, folds and pen marks lower confidence and are listed in quality flags

Optional context

The indication (first biopsy, active surveillance, prior treatment) and clinical history can go in clinical_context; they shape the guidance and the draft report.

The result

One JSON result per case, organised in sections, with per-core and case-level values side by side.

Section	Content
Cancer detection	Per core: carcinoma probability, heatmap, tumour length in mm and % of core involved
Gleason grading	Gleason pattern map, Gleason score, ISUP grade group 1–5, % pattern 4/5 and cribriform presence per core and per case
Recurrence risk	Biochemical-recurrence risk score and time-to-event estimate
Draft report	Opt-in (options.report): an English narrative that interprets the result for the reader (findings, impression, limitations); every number is checked against the findings

Standards

• Grading: ISUP 2014 grade groups with the ISUP/GUPS 2019 recommendations for reporting % pattern 4 and cribriform growth.
• Reporting: CAP prostate needle-biopsy cancer protocol; WHO Classification of Tumours, 5th edition.
• DICOM: heatmaps as Parametric Map, measurements and grades in SR TID 1500, the report as Encapsulated PDF; FHIR R4 DiagnosticReport with one Observation per score, and IHE PaLM APSR for LIS integration.

Grading notes

The grade group follows from the Gleason score: 3+3 is group 1, 3+4 group 2, 4+3 group 3, any score of 8 group 4 and 9–10 group 5. Cores with grade group 2 or 3 also carry the percentage of pattern 4.