Cancer Patient Management System - Architecture

System Overview

A comprehensive oncology care platform managing 8 real cancer patient cases with full medication protocols, role-based access, clinical alerting, and automated testing.

Architecture Layers

+--------------------------------------------------+
| PRESENTATION LAYER (Browser)                      |
| Single-page HTML/CSS/JS application               |
| Role-based UI (Admin/Doctor/Nurse/Patient views)  |
+--------------------------------------------------+
           | HTTP/JSON API
+--------------------------------------------------+
| API LAYER (Python HTTP Server)                    |
| REST-like endpoints, session authentication       |
| Role-based access control, audit logging          |
+--------------------------------------------------+
           | SQLite queries
+--------------------------------------------------+
| DATA LAYER (SQLite with WAL)                      |
| 11 tables, 10 indexes, foreign keys               |
| 8 patient cases, 35+ medications, 8 alerts        |
+--------------------------------------------------+

Database Schema (11 Tables)

Cancer Cases Included

Role-Based Access Control

Clinical Alert System

Automatic alerts trigger on:

• Critical: Febrile neutropenia (ANC <500 + temp >=38.3C), anaphylaxis, sepsis

• High: SpO2 <94%, cumulative neuropathy Grade >=2, pulmonary symptoms on bleomycin

• Medium: Rising tumor markers, LVEF monitoring overdue, bone pain >=7/10

• Low: Routine follow-up reminders, medication refill needed

API Endpoints (17 endpoints)

• POST /api/login - Authentication

• GET /api/patients - List patients (role-filtered)

• GET /api/patients/:id - Patient detail

• GET /api/patients/:id/medications - Patient medications

• GET /api/patients/:id/vitals - Vital signs

• GET /api/patients/:id/labs - Lab results

• GET /api/patients/:id/symptoms - Symptom reports

• GET /api/patients/:id/cycles - Treatment cycles

• GET /api/patients/:id/alerts - Patient alerts

• GET /api/alerts - All alerts (filterable)

• GET /api/users - User list (admin/doctor)

• GET /api/nurse-assignments - Nurse-patient assignments

• GET /api/audit-log - Audit log (admin)

• GET /api/dashboard/stats - Dashboard statistics

• POST /api/vitals - Record vitals

• POST /api/symptoms - Report symptoms

• POST /api/alerts/acknowledge - Acknowledge alerts

Testing System

20 automated tests covering:

1. Database connectivity and data integrity

1. All 8 patient cases loaded correctly

1. Medication records with full clinical detail

1. Vital signs, labs, symptoms, cycles populated

1. Alert system with critical/high/medium/low severity

1. User accounts and role assignments

1. Nurse-patient alignment

1. Audit log tracking

1. Authentication enforcement

1. Cross-patient data aggregation

File Structure

CancerPatientSystem/
  database.py          - SQLite schema, connection, password hashing
  seed_data.py         - 8 cancer cases, 35+ medications, vitals, labs, alerts
  api_server.py        - HTTP API server with 17 endpoints
  cancer_patients.db   - SQLite database (auto-created)
  static/
    index.html         - Single-page application UI
  docs/
    SYSTEM_ARCHITECTURE.md  - This file
    README.md               - Setup and usage guide
  tests/
    test_runner.py     - Automated test suite

Deployment

1. No dependencies required (Python standard library only)

1. Run: python api_server.py

1. Opens browser at http://127.0.0.1:8080

1. Database auto-creates with demo data on first run

Data Sources

Patient cases based on published oncology protocols:

• TCHP for HER2+ breast cancer (NCCN guidelines)

• Concurrent chemoradiation + durvalumab for NSCLC (PACIFIC trial)

• FOLFOX + bevacizumab for colorectal (NO16966 trial)

• Cabazitaxel for mCRPC (TROPIC trial)

• 7+3 induction for AML (standard ELN protocol)

• ABVD for Hodgkin lymphoma (standard first-line)

• mFOLFIRINOX for pancreatic (PRODIGE trial)

• Olaparib maintenance for BRCA+ ovarian (SOLO-1 trial)