tutorial

Castration-Resistant Prostate Cancer Care Tech Platform Monitoring Guide 2026

"A comprehensive guide to monitoring digital health platforms supporting CRPC care, covering PSA doubling time tracking, HRR/BRCA testing workflows, PSMA PET scheduling, 177Lu-PSMA dosimetry, and cabazitaxel neutropenia automation."

Castration-Resistant Prostate Cancer Care Tech Platform Monitoring Guide 2026

Overview

Castration-resistant prostate cancer (CRPC) represents one of oncology's most complex treatment landscapes. Defined by biochemical or radiographic progression despite castrate testosterone levels below 50 ng/dL, CRPC spans two major clinical states: non-metastatic CRPC (nmCRPC) and metastatic CRPC (mCRPC). Each demands a distinct armamentarium — from next-generation androgen receptor pathway inhibitors like apalutamide, darolutamide, and enzalutamide for nmCRPC, to the expanded mCRPC toolkit encompassing enzalutamide, abiraterone acetate, cabazitaxel, PARP inhibitors (olaparib, rucaparib), and the radioligand therapy 177Lu-PSMA-617 (lutetium vipivotide tetraxetan).

The digital health platforms supporting CRPC care teams are correspondingly sophisticated. EHR-integrated oncology modules, genomic result routing engines, nuclear medicine scheduling systems, and hematology protocol automation tools must all function reliably for patients to receive timely, personalized therapy. When these systems falter — a missed PSA trending alert, a BRCA result stuck in routing, a 177Lu-PSMA dosimetry calculation that never surfaced — the downstream clinical consequences are severe. This guide describes how engineering and clinical informatics teams can instrument and monitor these platforms to ensure continuous, high-fidelity operation.

Care Technology Landscape

CRPC care platforms span several interconnected domains:

Oncology Information Systems (OIS) — Platforms such as Varian ARIA, Epic Beacon, and Elekta MOSAIQ orchestrate treatment planning, scheduling, and clinical documentation. For CRPC, they manage both systemic therapy orders (chemotherapy, androgen deprivation, PARP inhibitors) and, increasingly, radioligand therapy workflows.

Genomic and Molecular Testing Portals — Homologous recombination repair (HRR) and BRCA1/2 germline and somatic testing results arrive from external laboratories (Foundation Medicine, Tempus, Myriad myRisk) via HL7 ORU messages or FHIR DiagnosticReport resources. Internal routing logic determines whether results trigger oncologist review, genetic counseling referral, or automatic eligibility flagging for PARP inhibitor therapy.

Nuclear Medicine Scheduling and Dosimetry Systems — PSMA PET imaging (68Ga-PSMA-11, 18F-DCFPyL) is now standard for staging and restaging. 177Lu-PSMA-617 therapy requires specialized dosimetry platforms integrated with radiation safety management systems. Scheduling these studies across nuclear medicine, radiation oncology, and urology requires tight workflow coordination.

Hematology and Supportive Care Automation — Cabazitaxel carries a significant neutropenia burden. Automated G-CSF protocol triggers — checking absolute neutrophil count (ANC) thresholds, chemotherapy cycle timing, and patient weight — must be embedded in the chemotherapy order verification workflow.

PSA Trending and Surveillance Dashboards — PSA doubling time (PSADT) is a key prognostic marker in nmCRPC, used to stratify patients for intensified therapy. Automated PSADT calculation engines must pull serial PSA values, apply validated logarithmic regression models, and surface results to the treating physician within the clinical workflow.

Key Monitoring Metrics

PSA Doubling Time Calculation Engine

Throughput and Latency

  • PSA value ingestion lag: time from lab resulted to PSADT recalculation (target: < 15 minutes)
  • Calculation queue depth: number of pending PSADT computations (alert threshold: > 50)
  • Dashboard rendering time: time for PSADT trend chart to load in EHR portal (target: < 3 seconds)

Data Quality

  • Patients with fewer than 3 PSA values in the calculation window (insufficient for reliable PSADT): flag for manual review
  • Outlier PSA values (> 3 SD from patient's own trend) triggering false PSADT acceleration: detect and queue for clinician confirmation
  • PSA assay methodology changes (different laboratory platforms): alert when assay type changes mid-series, as this invalidates PSADT continuity

Calculation Correctness

  • PSADT algorithm version pinning: alert when calculation engine version differs between production and staging
  • Results outside biologically plausible range (e.g., PSADT < 1 month or > 200 months): route to quality review queue

HRR/BRCA1/2 Germline Testing Result Routing

Message Ingestion

  • HL7 ORU or FHIR DiagnosticReport receipt confirmation rate (target: 99.9%)
  • Time from report finalization at external lab to receipt in routing engine (track by lab partner)
  • Failed/rejected message rate (target: < 0.1%)

Routing Accuracy

  • Results successfully matched to a patient encounter: track unmatch rate (target: < 0.5%)
  • Results routed to correct oncologist/genetic counselor within SLA (e.g., 24 hours for pathogenic variants)
  • Pathogenic BRCA1/2 results that triggered PARP inhibitor eligibility flag: confirm flag creation rate

Downstream Actions

  • Genetic counseling referral creation rate following pathogenic or likely pathogenic germline result
  • Time from BRCA result receipt to PARP inhibitor order entry for eligible mCRPC patients (target: < 5 business days)

PSMA PET Imaging Scheduling

Scheduling Efficiency

  • Median time from PSMA PET order to scheduled appointment (target: < 10 business days)
  • Order-to-scan completion rate: orders that never result in a completed scan (target: < 2%)
  • Scan cancellation rate with reason codes (patient, equipment, tracer supply)

Tracer Availability Integration

  • 68Ga or 18F-DCFPyL tracer inventory API availability (uptime target: 99.5%)
  • Failed tracer reservation requests: alert when scheduling system cannot confirm tracer availability

Report Turnaround

  • Time from scan completion to preliminary read (target: < 24 hours)
  • Time from preliminary read to attending radiologist attestation (target: < 48 hours)
  • PSMA PET reports triggering restaging workflow: confirm downstream staging update rate

177Lu-PSMA-617 Dosimetry Workflow

Session Scheduling

  • Time from Lu-PSMA therapy authorization to first infusion appointment (target: < 21 days)
  • Cycle adherence rate: patients receiving all 6 cycles within protocol-specified windows (± 1 week)
  • Cycle delay rate with reason codes (hematologic, renal, patient preference, logistics)

Dosimetry Calculation

  • SPECT/CT dosimetry image upload success rate (target: 100%)
  • Time from SPECT/CT acquisition to dosimetry report generation (target: < 48 hours)
  • Absorbed dose calculation completion before next cycle: SLA compliance rate
  • Radiation safety attestation completion by RSO before therapy: compliance rate

Renal Function Monitoring

  • Pre-cycle creatinine and eGFR retrieval from lab system: automated pull success rate
  • eGFR threshold alert (< 30 mL/min/1.73m²): alert-to-oncologist notification time (target: < 2 hours)

Cabazitaxel Neutropenia G-CSF Protocol Automation

ANC Monitoring

  • Pre-chemotherapy ANC retrieval from lab system: automated pull success rate
  • ANC < 1,500/mm³ detection-to-alert time (target: < 30 minutes)
  • G-CSF order auto-generation rate following protocol-defined ANC nadir (confirm with clinical pharmacy)

Cycle Safety Gates

  • Cabazitaxel orders blocked pending ANC: false-negative rate (orders that should have been blocked but were not)
  • Dose modification recommendations surfaced by the clinical decision support system: acceptance rate by oncologist

Febrile Neutropenia Surveillance

  • Patients on active cabazitaxel with fever > 38.3°C triggering emergency protocol alert: detection rate
  • Time from fever alert to oncology on-call notification: target < 10 minutes

Platform Setup

Unified Monitoring Stack Architecture

Deploy a centralized observability stack that aggregates signals from all CRPC care platforms:

# Example Prometheus scrape config for CRPC care platform endpoints
scrape_configs:
  - job_name: psa_calculation_engine
    static_configs:
      - targets: ['psa-engine.internal:9090']
    metrics_path: /metrics
    scrape_interval: 60s

  - job_name: genomic_routing_service
    static_configs:
      - targets: ['genomic-router.internal:9090']
    scrape_interval: 30s

  - job_name: nuclear_medicine_scheduler
    static_configs:
      - targets: ['nm-scheduler.internal:9090']
    scrape_interval: 60s

  - job_name: lu_psma_dosimetry
    static_configs:
      - targets: ['dosimetry-platform.internal:9090']
    scrape_interval: 120s

  - job_name: chemo_protocol_engine
    static_configs:
      - targets: ['chemo-cds.internal:9090']
    scrape_interval: 30s

Synthetic Transaction Monitoring

Instrument canary transactions for each critical workflow:

PSA PSADT Canary — Inject a synthetic PSA series for a test patient every 30 minutes; verify PSADT calculation completes and result appears in the dashboard within 15 minutes.

Genomic Result Routing Canary — Send a synthetic HL7 ORU message with a test pathogenic BRCA2 result every hour; verify routing engine acknowledges receipt, matches to the test patient, and creates the eligibility flag within 5 minutes.

Lu-PSMA Scheduling Canary — Simulate a therapy authorization request every 4 hours against the scheduling API; verify an appointment slot is proposed within the expected response time.

# Pseudocode: PSADT canary transaction
def run_psadt_canary():
    test_patient_id = "CANARY-CRPC-001"
    psa_series = inject_synthetic_psa_series(test_patient_id, values=[12.4, 18.6, 27.3])
    start = time.now()
    result = poll_for_psadt_result(test_patient_id, timeout=900)  # 15 min SLA
    latency = time.now() - start
    metrics.record("psadt_canary_latency_seconds", latency)
    if result is None:
        alert("CRPC PSADT canary: calculation not completed within SLA")
    elif not is_plausible_psadt(result.doubling_time_months):
        alert(f"CRPC PSADT canary: implausible result {result.doubling_time_months} months")

HL7/FHIR Interface Monitoring

For genomic result routing, instrument the interface engine (e.g., Mirth Connect, Rhapsody, Azure Health Data Services):

  • Message queue depth per sending lab partner: alert if queue depth exceeds 100 messages or age exceeds 60 minutes
  • Failed message counts per message type (ORU^R01, DiagnosticReport): alert on any non-zero failed count sustained for > 5 minutes
  • Message throughput rate: detect sudden drops that may indicate upstream lab connectivity issues

Dosimetry Platform Availability

177Lu-PSMA therapy is administered on a fixed 6-week cycle schedule. Dosimetry platform downtime between cycles blocks physics approval and delays subsequent infusions. Monitor:

  • Platform health endpoint: check every 5 minutes; alert immediately on unavailability
  • SPECT/CT image transfer pipeline: monitor DICOM transfer success rates from nuclear medicine scanners
  • Radiation safety officer (RSO) attestation portal: ensure UI is reachable during business hours; alert if unavailable > 15 minutes

Alerting Strategies

Severity Tiering

P1 — Immediate Clinical Impact (page on-call engineer + clinical informatics lead)

  • PSA trending engine down > 5 minutes
  • Genomic routing engine down; incoming lab results not being received
  • Cabazitaxel ANC safety gate not responding to pre-chemo verification requests
  • 177Lu-PSMA dosimetry platform unavailable within 48 hours of scheduled therapy infusion

P2 — Degraded Operation (alert engineering team, notify clinical informatics within 1 hour)

  • PSA PSADT calculation queue depth > 100 or age > 30 minutes
  • Genomic result routing SLA breach: pathogenic result not routed within 24 hours
  • PSMA PET scheduling API latency > 10 seconds sustained > 5 minutes
  • Dosimetry report generation time exceeding 72-hour SLA

P3 — Data Quality / Operational (ticket creation, next-business-day review)

  • Unmatched genomic results rate > 1% over a rolling 24-hour window
  • PSA outlier detection rate > 5% of new values (may indicate assay change or data migration issue)
  • Cabazitaxel dose modification recommendation acceptance rate < 70% (may indicate CDS fatigue)

Alert Routing

Route P1 alerts to PagerDuty with on-call escalation policy. Include in the alert payload:

  • Affected patient count (if determinable without PHI exposure)
  • Last successful transaction timestamp
  • Direct link to runbook for the affected system

Route P2 alerts to a clinical informatics Slack channel with a 1-hour acknowledgment SLA.

Route P3 alerts to a JIRA oncology informatics queue for sprint planning.

Runbook Requirements

Maintain runbooks for each P1 scenario. Each runbook must specify:

  1. Immediate mitigation step (e.g., restart service, fail over to secondary)
  2. Patient impact assessment procedure
  3. Clinical notification protocol (who to call, what to say)
  4. Post-incident review trigger criteria

For the genomic routing runbook, include a manual re-processing procedure for messages that failed during the outage window, with audit trail requirements for CLIA compliance.

Conclusion

CRPC care technology platforms carry a uniquely high clinical stakes burden. The precision of therapy selection — distinguishing nmCRPC patients who will benefit from apalutamide from mCRPC patients who qualify for PARP inhibition or 177Lu-PSMA-617 — depends on genomic results, imaging workflows, and calculation engines operating reliably around the clock.

Monitoring these systems requires more than uptime checks. Engineering teams must instrument the full clinical workflow: from PSA ingestion to PSADT dashboard rendering, from genomic report receipt to PARP inhibitor eligibility flagging, from PSMA PET order entry to dosimetry approval, and from cabazitaxel order entry to G-CSF protocol automation. Synthetic canary transactions, HL7/FHIR interface queue monitoring, and tiered alert routing tied to clinical impact severity create the observability foundation that allows institutions to catch and resolve platform failures before they propagate to patient care.

Investing in this monitoring infrastructure is not a technical nicety — it is a clinical safety requirement for oncology programs managing one of the most therapeutically complex cancers in modern medicine.

Monitor your app with Vigilmon

Free plan — 5 monitors, no credit card required. Up and running in 60 seconds.

Start free →
← Back to all articles